Authors - Palak Patel, Chintan Shah, Premal Patel Abstract - The use of Internet of Things (IoT) in healthcare systems the world has also measured substantial improvement in patient care by offering real time monitoring and diagnosis. But growing connectivity means that sensitive medical data is also vulnerable to numerous security threats like man-in-the-middle (MITM) attacks, key theft, and device cloning, among many others. These traditional methods for authentication, based on either passwords or static keys, are especially susceptible to such cyber-attacks; thus, they are ineffective in protecting healthcare IoT environments characterized by dynamic threat landscapes. This research article presents a new dynamic key-based authentication scheme for healthcare Internet of Things (IoT) networks. In the suggested framework, mutual authentication between each of the healthcare server and the physical IoT client device is used to protect the sensitive data of the patients in transit, along with providing real-time secure transmission. By focusing on how cryptographic keys develop, the framework addresses current security threats, including key theft and MITM attacks. Not only does this method keep computational and communication costs low; it is also lightweight enough for resource-constrained medical IoT devices using a lightweight authentication protocol. This system is efficient with respect to data size between registration and authentication phases while performing better than all other systems. It also minimizes latency and bandwidth usage by improving the caching layer. The scheme is also computationally cheap, making it feasible in resource-poor healthcare settings. The security analysis tools confirm that the suggested architecture is resistant to ordinary attacks such as MITM and replay attacks while incurring low overhead.
Authors - Abhijeet Kumar, Yash Shekhawat, Rahul Kumar, Naresh K Abstract - Real-time sign language recognition depends on an enhanced CNN-LSTM architecture which uses ASL training data. A new preprocessing approach boosts image resolution to 46×46 pixels thus enhancing the recognition precision. Numerous frames enter the system which enables a deep learning model to analyze spatial and temporal features to identify different hand signals in real time. The system integrates with the Flask-React-based frontend which allows real-time predictions through webcam interfaces in order to support practical field use. The system translates acknowledged signs into speech audio through Text-to-Speech APIs which drives inclusion between hearing-impaired users and people who do not use sign language. The system’s effectiveness was validated through experimental testing which reached 95% accuracy. The system's development will advance by implementing support for local sign languages with mobile platform integration.
Authors - Vitika Soni, Sakshi Chauhan, Harsho Mohan Chattoraj, Varun Dutt Abstract - There have been few direct comparisons in the literature between the emotional and autonomic impact of virtual reality (VR) and traditional and visual-narrative media, although VR has the potential to create vivid engagement. This study bridges this gap by using the culturally relevant narrative of King Harishchandra from the Indian Knowledge Systems (IKS) corpus to examine the impact of various storytelling media—text reading, comics, and virtual reality (VR)—on emotional and physiological engagement. We measured physiological arousal through Heart Rate Variability (HRV) indices, including RMSSD (parasympathetic modulation) and LF/HF ratio (sympathovagal balance), and emotional reactions through the Positive and Negative Affect Schedule (PANAS) in 30 participants (N = 10 per group). Comics performed better than text and virtual reality in terms of ratings of positive affect. While the VR and text groups reported significantly higher RMSSD values, reflecting more parasympathetic activity, comics registered the lowest LF/HF ratios, reflecting smoother cognitive-emotional processing and greater autonomic balance. The findings indicate that comics, despite being frequently neglected, can stimulate more stable emotional engagement than virtual reality immersion, which has implications for future application in therapeutic design, education, and culturally-based digital media. By emphasizing the importance of narrative coherence and cognitive manageability in shaping user experience, this research contributes to digital storytelling, affective computing, and media psychology.
Authors - Zaibunnisa L.H. Malik, Amena Niyaz Ahmed Malik, Pooja Raundale Abstract - Autistic Spectrum Disorder (ASD) refers to a group of developmental disorders that affect the nervous system, leading to challenges in social interaction, communication, and behavior. The severity of ASD symptoms can vary widely, ranging from mild to severe. Diagnosing and predicting ASD with high accuracy requires the use of advanced machine learning models. However, one of the major challenges in building such models is the availability of sufficient data. Open-source datasets often have a limited number of instances, which may not be enough to train robust models that can generalize well to new, unseen data. To overcome this limitation, it is essential to augment the dataset with additional, synthetically generated instances. In this context, techniques like corGAN (Conditional Generative Adversarial Networks) are employed. This comprehensive dataset is then used to train a machine learning model, which can more effectively predict ASD. The synthetic data ensures that the model has access to a richer, more varied set of information, ultimately leading to better performance and more accurate predictions for diagnosing and understanding ASD. We will also apply SMOTHE and Adaptive_Synthetic on GAN data, and prove that SMOTHE on GAN data gave a better distribution than Adaptive_Synthetic on GAN data.
Authors - Amitabha Bhattacharyya, Ayush Misra, Sourkarjya Kundu Abstract - The Relational Database was widely used in Industry and academics for software development in the early 90's and it is continuing now even starting from Oracle 7.1.3 version. But with increasing complexity of data, unstructured data or semi structured data capturing, research has shifted the paradigm from SQL to NoSql. Variety of NoSql databases has emerged and could not sustain in the market because of their own limitations. The Relational Database System has complex join operation and is costly compared to NoSql. 80% market share is grabbed by Oracle for any new software development in 2005 and later on researcher started inventing NoSql with low cost, open source and it’s quite challenging. In our IEEE 2020 we did survey different types of Graph databases and also showd the path on how to reengineer SQL to NoSql in the Springer publication. This paper aims to provide insights about Nosql MongoDB, relational database Oracle or MySQL, re-engineering methodologies study from sql to NoSql and also how much faster is SQL vs NoSql.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Devashree Kute, Anjali Naik Abstract - This study introduces a resource-efficient method for generating artwork in the traditional Madhubani style using Stable Diffusion XL (SDXL). Using the low rank adaptation (LoRA) technique, the model is fine-tuned with culturally relevant prompts and stylistic guidance to emulate the distinct characteristics of this Indian folk art form. To improve inclusivity and accessibility, prompt multilingual support is incorporated, covering Hindi, Bengali, Telugu, and English, through an automated translation mechanism, which retains the keyword to ensure consistency between languages. The training process is optimized for standard consumer-grade GPUs utilizing FP16 precision, CPU memory offloading, and a fixed Variational Autoencoder (VAE), enabling stable 1024×1024 image generation. For evaluation, a Contrastive Language Image Pretraining (CLIP) based scoring method is employed to assess the semantic alignment between prompts and generated images. The findings indicate that the style remains remarkably consistent across languages, demonstrating that even minimal training can enable AI to effectively capture and preserve traditional art forms within digital media.
Authors - Sonali Patil, Adwait Jadhav, Sahil Bhavsar, Pradnya Kamble, Arya Tandale Abstract - This research introduces the innovative AI-driven dashboard, leveraging Transfer Learning on Efficient Net and Retrieval Augmented Generation (RAG) for recommendation generation. The research is focused on minimizing the losses incurred due to diseases thus increasing the crop yield to meet the increasing demand. Our system incorporates the pre-trained Efficient Net model to classify the leaf images into various disease categories while LLAMA 3.1 8B LLM model is used to generate the remedy insights. Our methodology not only tackles the issue of hallucinations as well as correctness of the information which are very common in LLM response generation, but provides a personalized remedy plan for the farmer incorporating climatic conditions such as Temperature and Humidity as well. The suggested methodology not only aims at detecting diseases at early stages, but also at securing the necessary food supply, reducing the amount of pesticides used and promoting eco-friendly way of cultivation. Looking at the future, the project visions of increasing the use of LLMs in the agriculture industry by continuously upgrading RAG as well as the supplied documents to maintain up to date responses.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Athira T, Abhishek H, Deepak Gupta, Shobhana Palat Madhavan Abstract - With growing environmental awareness, consumers are drawn to sustainable offerings. Greenwashing — false environmental claims—has made consumers skeptical and impacted purchasing behaviour. Understanding greenwashing perception is vital to studying sustainable consumer behaviour, especially in India where research is limited. This paper explores greenwashing and sustainable consumer choices through a quantitative research design, using data from 223 Indian consumers analyzed via SEM and Mediated SEM in Stata. Employing the Value-Belief-Norm (VBN) Theory and the Theory of Planned Behaviour (TPB), the study examines the role played by environmental knowledge, green skepticism, social influence, and perceived moral obligation in sustainable consumption and green purchase intention, with greenwashing perception acting as a mediator. The findings suggest that perceived moral obligations are the strongest driver for greenwashing perceptions, as well as for green purchase intentions and sustainable consumption behaviour. Environmental knowledge enhances green-washing perception and sustainable behaviour. Green skepticism affects sustainable choices indirectly via greenwashing perception. Social influence affects greenwashing perception in a negative manner but sustainable consumption and purchase intention in a positive manner. These insights can help shape strategies that build consumer trust and promote genuine sustainable practices.
Tuesday August 25, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Keyur Kurani, Vaibhavi Machchhar Abstract - Ransomware is a rapidly evolving cyber threat with the potential to cause significant financial and operational disruption across industries. This paper analyzes the 2017 WannaCry attack, which exploited a Microsoft Windows vulnerability to infect over 200,000 systems globally, severely affecting sectors such as healthcare, including the UK’s NHS.The study explores ransomware’s progression from simple attacks to advanced, AI-powered variants. These modern forms use real-time analysis, intelligent target selection, and adaptive evasion techniques to bypass traditional defenses and enhance social engineering tactics.To counter these threats, organizations must adopt proactive strategies, including AI-driven cybersecurity solutions, behavior-based anomaly detection, and targeted employee training. Understanding the AI–ransomware link is essential for building resilient defenses against future cyberattacks.
Authors - Purvee Agrawal, Sanika Bhosale, Viraj Kakade, Vishal Jaiswal, Sarthak Baraliya Abstract - This paper presents the design and implementation of a Web Application Firewall (WAF) using machine learning models to effectively detect and mitigate three prominent web security threats: Distributed Denial-of-Service (DDoS), SQL injection, and Cross-Site Scripting (XSS). The proposed system leverages separate machine learning models for each attack type, optimizing detection by focusing on specific features unique to each threat. By analyzing traffic behavior, request payloads, and input structures, the WAF ensures high accuracy in identifying and blocking malicious activities. This multi-model approach significantly reduces false positives and enhances real-time protection. The solution is scalable and can adapt to evolving attack patterns, providing robust security for modern web applications and critical infrastructure.
Authors - Amulya Naik, Pallavi Dhaded, Shireesh Hakki, Satish Chikkamath, Suneeta V Budihal, Sujata Kotabagi Abstract - A text processing framework that applies Encoder-Decoder architecture with attention mechanism functions as the main focus of this research for resolving Natural Language Processing predictive problems. The research first outlines base technologies along with methodologies and frameworks required to build the system design. Detailed analysis of the dataset happens at this phase through observing dataset structure and calculating statistical summaries to detect missing or duplicate values. Better understanding of the dataset by using descriptive analytics to identify potential problems which leads them to improve the dataset. Data cleaning serves multiple functions during the process by eliminating unneeded columns together with missing value management and text normalization methods. The normalization procedure entails converting text into either upper or lowercase format and executes tag stripping alongside URL replacement and shorthand elimination and emoji and contraction removal. The processing begins after tokenization divides the text into segments. After cleaning the data the input and output components get separated while padding is used to maintain consistent dimensional structure.The model base incorporates an Encoder-Decoder framework combined with attention functionality while implementing a BiLSTM network. Through this specific model configuration both past and future inputs can be read contextually which boosts the prediction accuracy. The designed model produces 90.23 percent achievement in accuracy which highlights its strong capability in processing intricate NLP operations.
Authors - Abirami K, Megha Nayanar, Shobhana Palat Madhavan, Deepak Gupta Abstract - This study explores the impact of AI anxiety on career anxiety and career uncertainty, using Self-Determination Theory as its theoretical framework. As Artificial Intelligence (AI) continues to reshape industries, it presents both opportunities and challenges in career related factors. AI anxiety, driven by concerns over job security and skill obsolescence, affects individuals' confidence in their career choices and decision-making processes. As a result, many individuals may experience uncertainty about their professional future. By analyzing data from 237 respondents across India, this study identifies AI anxiety as a significant factor influencing career anxiety and career uncertainty. The findings reveal that individuals with higher AI anxiety are more likely to experience higher anxiety and indecisiveness in their career. However, the overall life satisfaction helps in lowering career anxiety by providing individuals with a higher sense of well-being that counteracts the anxiety. In contrast, factors such as relatedness, competence, and AI resilience do not show significant influence on either career anxiety or career uncertainty. This study enhances the understanding of how AI anxiety shapes career-related concerns, offering insights into how individuals navigate career decisions in an AI-driven world.
Tuesday August 25, 2026 12:30pm - 2:30pm IST Virtual Room DGOA, India
Authors - Shreya Ratagal, Parvati Navalur, Amruni Joshi, Satish Chikkamath, Sujata Kotabagi Abstract - With the instant advancement of generative artificial intelligence, this study investigates the domain of text-to-image generation, concentrating on the utilization of the Stable Diffusion model. The research examines the creation of visual content from written descriptions by leveraging sophisticated neural network architectures and underscores the importance of Natural Language Processing (NLP) in producing high-quality results. An extensive analysis was performed, integrating both qualitative and quantitative assessments to evaluate the model’s performance, scalability, and ability to adapt to various inputs. The study emphasizes potential uses in creative content production, virtual environments, and educational resources while tackling ethical issues to promote responsible AI practices. The results highlight the revolutionary effects of text-to-image generation in transforming the process of visual content creation.
Authors - Nishant Sharma, Mohit Mahlawat, Mohit Sharma, Gagandeep Singh, Ayush Kumar Singh, Kamlesh Sharma Abstract - An Early Warning System (EWS) utilizing Internet of Things (IoT) technology represents a transformative approach to disaster prevention and management. By leveraging interconnected devices, sensors, and real-time data transmission, IoT-based EWS enhances the ability to detect potential hazards—such as natural disasters, industrial failures, or environmental threats—at their earliest stages. These systems enable timely alerts and response strategies, minimizing risks to human life, infrastructure, and ecosystems. IoT technology plays a crucial role in gathering precise, real-time data from various sources, including seismic sensors, weather stations, water levels, and air quality monitors. This data is then transmitted to centralized platforms for analysis, allowing authorities and stakeholders to predict, assess, and act swiftly before a disaster strikes. With cloud computing and AI integration, IoT-enabled EWS can also deliver highly accurate forecasts and automated decision-making, further enhancing disaster resilience. As the world faces increasing threats from climate change, environmental degradation, and urbanization, IoT-based Early Warning Systems are becoming essential tools for safeguarding communities, enhancing preparedness, and ensuring a more resilient future
Authors - Jigme Nidup, Adithya Gattadi, Naresh K Abstract - Many serious health conditions, such as atrial fibrillation (AF), neuropathy, muscle disorders, and sleep-related neurological diseases, often go undiagnosed until complications arise. To address these challenges, this paper presents an advanced health monitoring system that integrates a multifunctional 4-in-1 electrogram sensor capable of measuring muscle activity using Electromyography (EMG), eye movement using Electrooculography (EOG), brain activity using Electroencephalography (EEG), and heart rhythm using Electrocardiography (ECG), along with a body temperature sensor, into a compact and wearable device at low cost. The device leverages the ESP32 Wi-Fi module to process and enable seamless data transmission to a Message Queuing Telemetry Transport (MQTT) cloud platform, ensuring secure, efficient, and scalable storage and analysis of collected health data. The system uses multiple pre-trained CNN models, each specialized in detecting specific diseases. Tests show an average accuracy of 90.3% making it a cost effective and efficient solution.
Authors - Harsha M R, Jyotiradhitya Kallimani, Nitishgouda Patil, Tohid Bijalikhan, Satish Chikkamath, Suneeta V. Budihal, Sujata S. Kotabagi Abstract - In India what does it take to go PRO in football? Players from their childhood shred sweat,blood and their precious time. And in sport we know that the next-gen superstar is guaranteed to start off his career from local and youth leagues. And honestly speaking these leagues do not offer the resources to invest. Resources in the sense that include technical skills, opponent player data, and event data(player stats). Currently, event data is mostly collected manually by human individuals, who gather data in several steps and through numerous persons involved. And this manually collecting data requires a lot of human resources and requires multiple checks and for that reason collection data is not practical in local or youth leagues. And this process takes a lot of time. So Automatic event detection could provide event data faster. which players can take use to analyse players and their own performance. And through which scouts would be able to ensure no player is missed or overlooked.
Authors - Arushi Madaan, Sunita Garhwal, Anu Bajaj Abstract - Women today are most likely to be experiencing Polycystic Ovary Syndrome (PCOS), a hormonal imbalance disorder. This disorder mostly affects women’s ovaries, where a large number of tiny fluid-filled sacs called cysts—also referred to as follicles—form around the ovary’s periphery. The exact root cause of PCOS is still unknown despite advances in science. Using ultrasound (US) scans to identify numerous follicles is an efficient way to diagnose PCOS early and schedule treatment. The primary purpose of this article is to determine whether or not a woman has PCOS or not without supervision from a physician. In this work, we provide a deep learning (DL) method based on transfer learning for PCOS classification using US ovarian images, with the goal of improving diagnostic efficiency and precision. InceptionV3 and ResNet50 models, which had accuracy rates of 99.68% and 97.5%, respectively, were used for this research. The study’s findings show that, as compared to conventional machine learning (ML) techniques, transfer learning-based classification performs better in PCOS variant classification. This study aims to accurately diagnose PCOS in patients and use our proposed model to treat PCOS. Gynaecologists and other medical professionals can benefit from our model’s ability to provide a prompt, reliable, and correct response.
Tuesday August 25, 2026 12:30pm - 2:30pm IST Virtual Room DGOA, India
Authors - Sushma Vispute Priya Surana, Shubhangi Vairagar, Sujit Shaha, Omkar Shinde, Sameer Sambhare, Krushna Salbande Abstract - Autism Spectrum Disorder (ASD) is a complex neurodevelopmental condition affecting social interaction, communication, and behavior. Autism is the third most common developmental disorder in the world. In India, the prevalence of autism is increasing and is estimated to be around 1 in 68 children. With its rising prevalence, early detection is crucial for timely intervention. This paper serves as both a review and a research study. The review explores existing ASD detection methodologies, highlighting machine learning approaches such as multinomial logistic regression (MLR), support vector machines (SVM), and convolutional neural networks (CNN), along with tools like eye-tracking and EEG analysis. The research component applies machine learning models—including Logistic Regression, Decision Tree, Random Forest, SVM, KNeighbors, Naive Bayes, and Neural Networks—on AQ10 survey data (1054 samples, 19 features) to evaluate their effectiveness. SVM achieved the highest accuracy. Further analysis examined the necessity of all 10 AQ10 questions, revealing that AQ4 and AQ10 contribute the least to predictive accuracy. Heatmap analysis confirmed weak correlations with the total ASD score. These findings suggest that refining ASD screening tools by removing less informative questions can improve efficiency while maintaining diagnostic reliability.
Authors - Vijayalakshmi S Katti, Usha J Abstract - Pests, pose a pervasive threat to agriculture on a national scale. Their voracious feeding on plant roots results in diminished crop yields, leading to economic losses for farmers and potential food security challenges. The integration of advanced sensor technologies and deep learning offers a promising avenue to address the impact of Pests, enabling timely and accurate detection. This, in turn, allows for targeted and efficient pest management strategies, mitigating the widespread repercussions of infestations and fostering sustainable agricultural practices on a national level. This paper explores the transformative synergy between sensor technologies and deep learning techniques for the identification and density detection of pests in agriculture. Traditional methods face limitations, prompting a shift towards advanced technologies. We survey the landscape of sensor technologies, including image sensors, acoustic sensors, and soil sensors, highlighting their real-time, high-dimensional data contribution. Integration with deep learning models, such as Convolutional Neural Networks and Recurrent Neural Networks, offers a precise and adaptive approach to pest management. The potential impact of this integration is substantial, promising increased crop yields, reduced reliance on broad-spectrum pesticides, and improved environmental sustainability. The review underscores the importance of continuous adaptation and scalability, setting the stage for a future where technology plays a pivotal role in ensuring the health and productivity of agricultural landscapes.
Authors - Ashwini Jarali, Sanskruti Lad, Snehal Kavathekar, Prajwal Lalpotu, Shreya Jadhav Abstract - Potholes on roads significantly impact safety and road infrastructure, leading to accidents and increased vehicle damage. Timely detection and repair are crucial to address these issues effectively. This paper presents an AI-based system for automated pothole detection and reporting, aimed at improving pothole management for road maintenance authorities. The system uses the YOLO (You Only Look Once) object detection model to accurately identify potholes in real-time road imagery, combined with GPS for precise localization. Detected potholes are automatically reported to the relevant authorities via email, ensuring swift corrective action. The YOLO model is trained on a diverse dataset of pothole images, achieving high detection accuracy across various pothole sizes and shapes. Additionally, the system tracks the status of reported potholes to ensure repairs are completed. This solution enhances road safety and reduces manual effort, providing a comprehensive approach to modern road maintenance.
Authors - Harshit Kadam, V Harshavardhan, Shrigouri S G, Bhavana Gadagin, Nalini Iyer, Prabha Nissimagoudar Abstract - One of the key challenges for blind spot detection systems is the ability to detect and track objects in irregularly shaped regions. This problem becomes much more severe in addition to considering different vehicle velocities, motions due to other objects, and many different environmental conditions. Ordinary systems have fixed parameters for operating conditions, which either are slow or may not adapt to changeable driving scenarios. The proposed solution is an adaptive continuous monitoring approach that provides the defined polygons with the ability to report on any encroachments while giving a proximity risk based on context data. This real-time adaptability allows the system to provide accurate and timely notifications to the driver, thereby increasing the safety of critical events such as lane changes, parking maneuvers, or heavy traffic situations where blind spot threats are most prevalent. Initial results show that this adaptable system can outperform traditional precision and time response methods, improving overall safety while driving.
Tuesday August 25, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - Sharim Iqbal, Purnima Ahirao, Deepti Patole Abstract - QR code payment systems have become increasingly popular thanks to their speed, simplicity, and convenience. However, as their usage grows, so do concerns around security—issues like tampering, spoofing, and man-in-the-middle attacks are becoming more common. To address these vulnerabilities, this paper introduces a novel approach to securing QR-based transactions using visual cryptography. Visual cryptography works by splitting an image into multiple shares, which individually reveal nothing but can reconstruct the original image when overlaid—without the need for complex decryption algorithms. This research proposes a secure QR payment system that leverages visual cryptography to enhance data integrity, prevent fraud, and strengthen the overall security of mobile payments. The study covers current security challenges, outlines the proposed system architecture, implementation details, and evaluates its performance.
Authors - A.Akshaya, G.Veera yasaswini, P.Akshith, K.Mahimanusha, M.TanviSahasra, Sushmarani Abstract - Sentiment analysis is among the primary natural language processing (NLP) tasks and is widely utilized for extracting emotions and sentiments from text corpora. This paper proposes a comprehensive sentiment analysis approach for movie reviews based on Word2Vec, TextBlob, VADER, and Gated Recurrent Units (GRU). Word2Vec is employed for word embeddings to extract semantic word relationships for improved feature representation. TextBlob and VADER are implemented as lexicon-based sentiment analysis tools, for which TextBlob is interested in polarity and subjectivity and VADER is engineered for short texts with clear-cut sentiment indications. Besides, deep learning architecture in the form of GRU is employed for extracting long dependencies and context associations between words of text corpora for enhanced sentiment classification. Methods are experimented and contrasted on the basis of a benchmark IMDB dataset with reference to accuracy, precision, recall, and F1 score. Experimental findings substantiate that sentiment handling by deep learning-based approaches, i.e., GRU via Word2Vec embeddings, is better than traditional lexicon based approaches. The work provides insights to NLP-based opinion mining researchers and practitioners regarding the merit of utilizing hybrid approaches towards sentiment classification.
Tuesday August 25, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - V M Aparanji, Chaithanya C K, Nanditha N, Poornima H S, Shreya A N Abstract - Face recognition is a vital biometric technology with growing applications in security, access control, and automation. Despite challenges such as lighting variation, facial expressions, pose angles, and aging effects, Support Vector Machine (SVM) has proven effective in modeling and classifying facial features due to its ability to construct optimal decision boundaries in high-dimensional spaces. In this study, SVM was applied to a face recognition system for smart lock operations, yielding strong performance metrics of 89.9% accuracy, 89.7% precision, 89.5% recall and an F1 score of 89.7%. These results demonstrate the capability of SVM to effectively manage facial variability while maintaining high accuracy in face recognition tasks.
Authors - Ajay Talele, Amruta Mankawade, Aryan Sutar, Nishit Shelar, Urvesh Somwanshi, Anushka sonde, Shiv Sagar Singh, Sharvari Savardekar, Shivanand Satao, Shridhar Sarda, Raj Bapat Abstract - A Medicine Recommendation System intended to use user-provided symptoms to identify possible diseases and provide personalized suggestions for safety measures, diets, drugs, and exercise regimens. The system predicts symptoms using Natural Language Processing (NLP) and Fuzzy Matching, guaranteeing accurate identification even in the presence of noisy inputs. It makes predictions about likely diseases and obtains overarching information for each by comparing extracted symptoms with a disease- symptom dataset. The system, which was developed with the Flask framework, provides an intuitive online interface for smooth communication. This initiative aims to direct users toward informed medical treatment by showcasing potential in early disease identification. Future research will concentrate on improving accessibility more broadly and integrating healthcare data in real-time.
Tuesday August 25, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - Praniv Warungshe, Siddhi Desale, Sushant Mhatre, Rujata Chaudhari, Saylee Lapalikar Abstract - After graduation, alumni often face difficulties in managing important academic documents such as Leaving Certificates (LCs), marksheets, Letters of Recommendation (LORs), and convocation updates. Traditional processes are manual, time-consuming, and lack real-time tracking, leading to delays and repeated campus visits. To address these challenges, a cross-platform application was developed using React Native to simplify and digitize post-graduation work-flows. The application enables alumni to request, verify, and correct documents through a single digital platform with real-time status updates. LCs and marksheets are available on-screen, making the verification process faster and more convenient. Convocation details can also be managed easily within the app. Faculty benefit from tools to handle LORs and other academic tasks efficiently, while administrators use a centralized dashboard to track applications, generate reports, and respond to urgent requests quickly. The cross-platform nature ensures seamless access across various devices, enhancing user convenience. By uniting all stakeholders on one platform, the system boosts transparency, reduces manual workload, and modernizes the overall process, offering an efficient and user-friendly solution to manage alumni post-graduation needs.
Authors - Abhishek A Joshi, Vasudhaika S, Sinchana Chindi, Kaushik Mallibhat, Satish Chikkamath Abstract - The proposed work aims to present a novel framework to classify the oral diseases through multimodal data consisting of images and symptoms. The outcome of the work helps towards early diagnosis of oral diseases. Oral diseases are neglected by most people in the initial stages due to lack of awareness, accessibility, and availability of dental care. Early diagnosis of oral disease is necessary, and early research focused on either the image of the affected area or just the textual description (symptom) to predict the illness; both are essential. The study employs multimodal approach and utilizes an image dataset comprising seven categories of affected areas, sourced from Kaggle. Additionally, a textual symptoms dataset was developed, consisting of 150 combinations for each illness. Confidence scores of both the model (image-classifier model and symptom-based illness prediction model) with the true label, a new dataset is generated and it is trained with logistic regression to get the final predicted class. Image classifier model achieved 81% of accuracy whereas symptom-based model 97%, resulting in final multimodal accuracy of surpassing both the model accuracies.
Tuesday August 25, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - Laxmi Sanjay Badiger, Keerti, Sumangala Basavaraj Donkanavar, Satish Chikkamath Abstract - Sentiment analysis is a vital task in NLP. That identifies the emotions in the text. Many studies concentrate only on the English language. There is an insufficient resource in Kannada language emotion analysis. This paper examines the sentiment in Kannada text using a manually prepared dataset. The datasets are divided into three classes positive, negative, and neutral. The processing techniques like context cleaning, tokenization, and sequence padding are used. The model uses RNN-LSTM which is efficient in handling the sequential data. The embedding layer is used to represent the words, the LSTM layer is used to get the context of the sentence, the Dropout layer is used to reduce the over fitting of the model and the dense layer is used to classify the sentiments into categories. The effectiveness of the model was measured using evaluation metrics like precision, recall and f1 score. By predicting sentiments for Kannada text, the paper also exhibits practical use of the model. This study demonstrates that LSTM-based models work well for sentiment analysis in Kannada. It also emphasizes the importance of creating and using manual datasets for low resource languages. The findings would be helpful to further research, and the out-comes can be directly applied in many areas, including social media monitoring, customer feedback analysis, and regional language processing.
Authors - Bhadouriya Khushi Mukeshsingh, Rajput Adityasingh Shashikantsingh, Parmar Smit Dharmeshkumar, Tiwari Prashant Dineshkumar, Soumya Kiran Prajapai, Nirav D. Mehta, Anwarul M. Haque Abstract - The accelerating proliferation of electronic devices has led to a surge in electronic waste (e-waste), presenting a significant environmental and resource management challenge. Conventional e-waste disposal practices are inadequate, often resulting in the release of hazardous substances and the loss of valuable materials. This paper explores a sustainable framework for e-waste reduction by leveraging advancements in power electronics and promoting standardization across electronic design and manufacturing. Key areas of focus include the adoption of fixed-type ports for power and data transfer, implementation of mandatory certification and testing standards for electronic components, and the design of modular and fixed PCBs to facilitate component reuse. Emphasis is placed on developing universal and multipoint-compatible components, particularly in the context of electric vehicle (EV) charging infrastructure, where interoperability can significantly reduce hardware redundancy. The integration of circuit protection mechanisms is proposed as a means to extend product lifespan and minimize failure-induced waste. Furthermore, strategies for the reuse and remanufacturing of components are examined as critical elements of a circular economy. By combining technical, regulatory, and design-driven approaches, this study outlines a comprehensive pathway toward reducing the environmental footprint of electronics and fostering sustainable innovation in power electronics.
Authors - Ayush Kayasth, Vrund Raval, Keren Khambhata, Nirali Nanavati Abstract - India leads the world in textile production and exports, but it also faces an increasing environmental problem: the country produces about 7,800 kilotons of textile waste a year, or 8.5% of the world's total. Even with dispersed efforts at recycling and reuse, progress is still hampered by the lack of a centralized, digitalized infrastructure for managing textile waste. In order to reduce waste in the Indian textile industry, our study proposes TextileLoop, a mobile application based on the ideas of the circular economy. The research employs a qualitative methodology, with focus group discussions conducted among key stakeholders in Surat India’s leading textile hub. These conversations uncovered important issues, such as small buyer networks, reliance on offline trade, and traditional players' reluctance to adopt new technologies. The results served as a guide for creating Textile Loop, a B2B platform with an MVC architecture developed with Flutter and Appwrite. In addition to providing educational materials and a collaborative environment for industry stakeholders, the application makes it easier to exchange excess textiles, faulty goods, and used machinery. Its integrated approach, which combines digital infrastructure with circular economy strategies, is what makes it innovative and the first circular economy based platform in India. Our platform aims to scale to a B2C and C2C model, allowing for greater engagement. With implications for both industry and policy, this work offers a workable plan for converting India's textile sector into a circular and sustainable ecosystem.
Tuesday August 25, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - Akshar Sodankoor, Avanish Shenoy, B Monish Moger, Mohnish Gowda, Prafullata Kiran Auradkar, Subramaniam Kalambur Abstract - Virtualization is essential for efficient resource utilization in cloud and development environments. With the growing adoption of gRPC as a Remote Procedure Call (RPC) framework, evaluating its performance across different virtualization technologies has become crucial. This work benchmarks the performance of four gRPC call types: unary, client-streaming, server-streaming and bi-streaming, across four lightweight virtualization technologies. Docker, gVisor, Firecracker and nanos unikernel. The analysis examines CPU utilization, memory utilization and network capabilities to provide a comprehensive comparison. The results show that docker delivers the best performance across all metrics. Firecracker shows comparable latency performance to docker, but consumes higher memory. Nanos unikernel exhibits CPU utilization similar to that of docker, but has the highest latencies in all cases except unary gRPC call. gVisor exhibits the lowest CPU utilization under heavier workloads and also has the lowest latencies for client-streaming and server-streaming gRPC calls.
Wednesday August 26, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Aryan Goyat, Aditya Maan, Vimmi Malhotra Abstract - Deep learning has transformed artificial intelligence and enabled major breakthroughs in applications like computer vision, natural language processing, medicine, cybersecurity, and robotics. Through the use of deep neural networks, it enables automatic feature learning and exceeds machine learning-based methods in accuracy and flexibility. Challenges including excessive computational expense, uninterpretable nature, and ethics are still major hurdles to its widespread application. This article discusses the development and applications of deep learning and presents new research directions that seek to overcome its limitations. Federated and decentralized learning methods improve security and privacy by enabling collaborative model training without raw data sharing. Explainable AI (XAI) techniques, including SHAP and LIME, enhance the interpretability of deep learning models, making their decision-making more transparent. In addition, energy-efficient deep learning methods, such as model pruning, quantization, and neural architecture search (NAS), are being designed to minimize computational and environmental expenses. The emergence of self supervised learning further minimizes dependence on labeled data, making deep learning more feasible across domains. Future developments will center on the fusion of deep learning with reinforcement learning, symbolic AI, and evolutionary algorithms to build more generalizable and efficient systems. These technologies will power the next wave of intelligent, ethical, and sustainable AI solutions.
Wednesday August 26, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Anita Agrawal, Ruhi Panjwani, Aditya Mallik Abstract - This paper explores the design and implementation of a multi-user, multi-access web application tailored specifically for automated weather stations (AWS). By examining real-world scenarios and user interactions, we identify key design considerations, including system performance, security, and scalability. The study aims to provide practical insights for developers to create efficient and user-friendly web applications that effectively handle large-scale weather data and cater to various user access levels.
Authors - Krishna Shirsath, Abdullah Ansari, Riyaz Memon, Phiroj Shaikh Abstract - Real-time collaboration is essential for modern software development, enabling developers to work together seamlessly from different locations. This paper presents DevTogether, a collaborative coding platform that facilitates efficient teamwork through live code editing, customizable collaboration sessions, integrated chat, a collaborative drawing board for design prototyping, and real-time video meetings powered by WebRTC. The platform incorporates an intelligent code assistant using the Gemini API, along with an autosave mechanism to ensure workflow continuity. Built using the MERN stack, DevTogether emphasizes scalability, low latency, and responsive performance while addressing synchronization conflicts and communication challenges. Experimental evaluations and simulated performance metrics underscore its effectiveness compared to similar platforms.
Authors - Arnav Shukla, Subhashree Choudhury, Logeshwaran R. Abstract - One of the key limitations of any crowdfunding platform that uses blockchain technology is the lack of transparency regarding fund usage by campaign creators after receiving donations. The current paper addresses this issue. For this investigation, we conducted a detailed examination of the relevant literature on decentralized applications (DApps) and the use of smart contracts to automate administrative tasks. We identified a significant gap in existing platforms: the ambiguity surrounding post-donation and how fund management can be unfair, which blockchain alone does not address effectively. Our findings highlight the strengths of blockchain security and automation capabilities, which ensure a safe and trustworthy environment for users. Our proposed solution integrates a decentralized voting model and a collusion prevention algorithm called EigenTrust to empower donors with a participatory role in decision-making processes, eliminating the need for centralized authority. In this study, we implement and evaluate a decentralized voting model that uses specific techniques to prevent any attacks or chances of misuse of powers that would then empower donors to participate in critical campaign decisions, enhancing trust and satisfaction by allowing them to verify the responsible use of their contributions. By reducing uncertainty around fund allocation, our model increases a more engaging and end-to-end secured donor experience, encouraging donations and supporting the long-term success of social crowdfunding projects. In all, this paper presents a novel approach to increase crowdfunding platforms using blockchain technology with a voting model paired with collusion prevention to address the issue.
Authors - P S Sree Harsha, Harshitha S, Ayush Sisodiya, M P Deepti, Sarasvathi V Abstract - Concerning the Conventional recruitment methods, which has many challenges such as the poor matching of the candidates to the right requirements, issues of transparency and issues of privacy in dealing with sensitive information. This leads to the hiring of candidates who are not qualified to meet their requirements and compromise the organizational data. To solve these problems, Human Resource Blockchain Intelligence Recommendation System (HRBIRS) is proposed to use Blockchain and Recommendation system technologies for getting an efficient recruitment process. Blockchain is evolving technology providing transparent and secure data. Privacy issues are resolved by decentralized architecture. The Hybrid recommendation system makes recruitment effective by determining the qualifications, skills and experience of the job seekers relevant to certain job recruitments posted by the HR. One of the most attractive features of HRBIRS is the peer approval and endorsement systems within the organization without bias to a particular candidate. This paper provides detailed information on the architectural design, implementation and its performance features and demonstrates how this system could be beneficial for recruitment processes throughout the organization by providing data security and enhancing the effectiveness of selecting the right candidate.
Authors - Arunangshu Giri, Dipanwita Chakrabarty, Manash Routray Abstract - The present study emphasized on how enrollment of loyalty programs at fuel retail outlets get enhanced through digital communication and intelligent promotional strategies. The study has evaluated the three major dimensions like participation intention and promotional efficiency to understand the efficacy of the loyalty programs organized by different fuel retail outlets. 454 Indian customers were interviewed over a three months period using a structured questionnaire. Cross-sectional descriptive research design was followed for the study. Both qualitative and quantitative analysis was done using NVivo and SPSS-28 software. The study revealed that customer participation in loyalty programs was highly influenced by flexible redemption options, digital reward system and exclusive benefits. Again, the study has shown how staff knowledge, promotional policies, customer loyalty and digital engagement influence promotional effectiveness. The study acknowledged the pivotal role of intelligent communication strategies in enrichment of customer adaptability towards digitized loyalty programs. Establishing a seamless communication between the customers and digital platforms along with effective staff training can be prudential for optimal customer engagement, sustainability and loyalty.
Authors - Lalithya Govardhan, Shalini M S, Gagan Deep P S Abstract - This work shows an extensive multimodal system of mood detection and customized playlist recommendation based on EEG, GSR, and face recognition. Brainwave activity for emotional evaluation is sensed by EEG electrodes, while GSR sensors provide skin conductance, heart rate variability, and temperature values for physiological behavior. Facial behavior with emotional facial expressions is determined through facial landmarks. Preprocessing entails Butterworth filters for EEG frequency bands, GSR data normalization, and facial feature extraction from a pretrained model (FER) to track eyebrow position, mouth curvature, and eye openness. EEG features are examined using frequency domain analysis, whereas GSR and facial features are classified using Random Forest. To increase precision, a fusion model aggregates predictions by weighted averaging or majority voting, with EEG assigned the greatest weight due to its high correlation with mood. After determining the emotional state, a suitable playlist is suggested: energetic songs for happiness, relaxing music for stress, comforting songs for sadness, and relaxing music for relaxation. This feature-based recommendation system enhances personalization through the use of features like tempo, genre, and mood to provide a dynamic and interactive listening experience for the user
Authors - A. Revathi, A. Sunidhar Reddy, Geetika Alapati, R. Pranay Abstract - This paper presents the performance of the grocery identification system concerning Telugu grocery items, considering both native and non-native speakers. Speech recognition for Telugu groceries presents a unique challenge due to variations in pronunciation, accent, and noise conditions. This study explores the implementation of a Gaussian mixture model (GMM) classifier in conjunction with rasta-perceptual linear prediction (RASTA-PLP) features to enhance the accuracy of Telugu grocery identification. Rasta-PLP effectively captures robust speech features by suppressing unwanted spectral variations, while GMM provides a probabilistic framework for classification. The proposed system is trained on a dataset comprising commonly used Telugu grocery names and evaluated under diverse acoustic environments. Experimental results demonstrate improved recognition performance, showcasing the effectiveness of RASTA-PLP in feature extraction and GMM in classification. This work contributes to developing efficient speech-based interfaces for regional language applications, facilitating voice-driven grocery identification systems. The recognition accuracy of the proposed system is approximately 99%, ensuring high reliability in real-world applications. This technology benefits society by aiding visually impaired individuals and non-Telugu speakers in grocery identification, enhancing accessibility and convenience. By enabling seamless voice-based interaction, promotes inclusivity and improves social equity through technological advancement.
Authors - A.Revathi, Reethikaa Vallinayagam, S. Sivaranjani, A. Deepthi Abstract - This research work introduces a system for identifying genuine speech and recorded (replay) speech through Mel-Frequency Cepstral Coefficients extraction and uses k-means clustering for classification purposes. Speech features obtained from various speakers undergo normalization procedures before receiving cluster assignment during training sessions. During the testing phase speaker identification depends on measuring the distance between input features against cluster centroids. The confusion matrix indicates system performance by showing correct genuine speech detection through high diagonal values yet exhibiting lower off-diagonal values to indicate possible attacks based on recorded speech. Auto-correlation together with cross-correlation serve to evaluate the similarities between speakers. Strong recognition of the same speaker is indicated by high auto-correlation values but weak cross-correlation values demonstrate effective differentiation between different speakers. The AVSpoof dataset serves as the experimental foundation because it includes ten recording subjects who are distributed between five male and five female speakers. The acceptance and accuracy evaluation for the system happens through testing samples which proves its ability to recognize genuine speech from recordings as well as identify distinct speakers properly.
Authors - Moushmee Milind Kuri, Ganesh Pathak Abstract - Cyberbullying is a growing concern across social media platforms, necessitating advanced detection mechanisms to mitigate its impact. Traditional machine learning models often struggle with understanding contextual dependencies and ensuring model interpretability. This paper proposes a hybrid deep learning approach that combines BERT and RoBERTA for feature extraction and Bidirectional Long Short-Term Memory (BiLSTM) networks for sequential dependency learning. To enhance interpretability, attention mechanisms such as Self-Attention and Bahdanau Attention are integrated, allowing the model to focus on crucial words contributing to classification. The proposed system aims to improve accuracy, scalability, and explainability while addressing key challenges in cyberbullying detection. This research lays the groundwork for developing more transparent and effective AI-driven moderation systems for online safety.
Authors - Devi V S, Durgalashmi C V Abstract - This study assesses the effectiveness of income tax deductions and exemptions in promoting savings and investments in India. The Indian government has implemented various tax incentives to encourage individuals to save and invest, including provisions under sections 80C, 80D, and others. These deductions and exemptions are designed to stimulate economic growth by fostering long-term financial planning among individuals. The research examines the impact of these provisions on individual taxpayers' behavior and their overall influence on savings and investment patterns. Through a comprehensive analysis of available data, the study identifies the key tax incentives that have led to increased savings in instruments such as Provident Funds, National Savings Certificates, and insurance products. Additionally, the research evaluates the extent to which these tax benefits contribute to fostering a culture of investment and financial security. The study concludes that while tax deductions and exemptions have provided some incentives for savings, their effectiveness is often limited by lack of awareness and financial literacy. To further promote savings and investments, the study recommends improvements in policy communication, accessibility, and the alignment of tax incentives with broader economic goals.
Authors - Akash Tibeli, Saroja V Siddamal, Suneeta V Budihal Abstract - The AHB to APB Bridge is crucial component in System-on-Chip (SoC) designs, Achieving efficient communication between the pipelined AHB bus and the non-pipelined APB bus. In the proposed work a AHB to APB bridge is built using a bridge architecture which enables to translate pipelined, burst-oriented, high speed AHB transactions into sequential, low-power APB transactions by maintaining synchronization and data integrity. It was developed with a FSM to manage transactions and pipelining to maintain efficiency. Verification was performed using a Universal Verification Methodology testbench environment through direct and random testcases of burst, single, sequential, non-sequential transactions. 80 testcases were tested to obtain a functional coverage of 88%.
Wednesday August 26, 2026 12:30pm - 2:30pm IST Virtual Room DGOA, India
Authors - Judy K George, Elizabeth Sherly Abstract - Convolutional Neural Networks are extensively employed in critical domains such as computer vision, medical imaging, and autonomous systems. Enhancing model interpretability by providing users with concise and context-relevant explanations of CNN decision making such as visualizing feature maps or saliency regions, enables a deeper understanding of the model’s internal representations and inference process. The proposed work presents a deep learning framework integrating a ResNet-based U-Net architecture with a Fixation Point Generator (FPG) to perform classification and saliency aware reconstruction on the hand-written dataset. The model leverages transfer learning by employing a pre-trained ResNet-18 as the encoder backbone, enabling robust feature extraction. A custom decoder reconstructs input images while a classification head predicts digit labels. To enhance model interpretability, a Fixation Point Generator predicts spatial attention maps (saliency maps) from high-level global features, highlighting regions of interest that influence model decisions. This implementation aims to bridge the gap between classification performance and model explainability, offering insights into the model’s focus areas through learned attention. The model got an accuracy of 98.44 on the Malayalam handwritten dataset, 97.81 on English handwritten dataset, and 99.56 on the MNIST dataset.
Authors - Meghali Kalyankar, Om Pratap Gajra, Prathamesh Vilas Sagvekar, Mehul lalit Sharma, Zoheir Shahid Shaikh Abstract - Deepfakes are an emerging threat to digital authenticity and security, hence a proper detection technique needs to be created in order to establish public study confidence. A thorough roadmap to the development of deepfake detection software has been provided in this paper, reviewing the state-of-the-art algorithms, such as XceptionNet, EfficientNet, and hybrid models integrating spatial and temporal analysis. It provides methodologies for implementation, data preprocessing, and software pipeline development, serving as a practical guide to researchers and developers. Theoretical study to application-oriented practice closes the gap in terms of bottom line development and adaptive detection systems addressing the growing menace of deepfake media.
Authors - PRANAY SAMAL, K R LOKESH KUMAR, CHAKILELA SAIRAJ, G VIDYA SRI, SUSHAMA RANI DUTTA, SUKLA SATAPATHY Abstract - Music plays a significant role in our day-to-day life, and selecting appropriate songs can enhance the experience. This paper describes an intelligent music recommendation system that applies machine learning to recognize what users prefer and recommend music that they will like. It incorporates various approaches, including considering user decisions and music attributes, to enhance suggestions. The system adapts based on user actions and refines recommendations over time. Findings indicate that this method provides easier and more precise music discovery. This paper emphasizes how technology can assist in providing a higher quality and better personalized music experience.
Authors - Arya Tripathi, Akash Mecwan Abstract - In recent years, the RISC-V architecture has emerged as a promising platform for embedded systems, offering flexibility and open-source accessibility. Consequently, the demand for secure communication in embedded devices, particularly within the Internet of Things (IoT) ecosystem, has driven the adoption of cryptographic algorithms. Integrating cryptographic functionalities into RISC-V architecture presents unique challenges, requiring innovative solutions to optimize performance and security. In response, the proposed design introduces an approach to address these challenges by incorporating a dedicated cryptoprocessor module into the RISC- V architecture specifically designed to handle encryption and decryption tasks efficiently. The cryptoprocessor module employs the Blowfish-64 algorithm to ensure robust security while lowering the computational overhead. Blowfish is a well-established symmetric-key block cipher known for its simplicity and efficiency. The compact design of the cryptoprocessor module significantly reduces resource utilization and execution time compared to existing implementations while preserving security and functionality. The design emphasizes low resource utilization, achieving a utilization rate of 34% (11,340 out of 33,216 available units) with an execution time of 160 ns. The implementation is carried out using Verilog HDL for the Cyclone II EP2C35F672C6 based FPGA.
Wednesday August 26, 2026 12:30pm - 2:30pm IST Virtual Room DGOA, India
Authors - Kirti Karande, Sujata Kadu, Deven Shah Abstract - Breadth-First Search (BFS) is a foundational graph traversal algorithm, it’s systematic layer-by-layer exploration of nodes makes it invaluable for a variety of domains, including transportation networks, social network analysis, and artificial intelligence. However, traditional BFS implementations face challenges when dealing with large-scale graphs due to memory limitations and inefficiencies in handling massive datasets. This project addresses these challenges by integrating BFS with a CSV-based data storage system, enabling efficient traversal of large graphs without relying on a traditional SQL database or requiring the entire graph to be loaded into memory. The graph data, comprising nodes and edges, is stored in CSV files, which act as lightweight and accessible storage. The implementation is memory-efficient due to the use of Pandas DataFrames for handling CSV data and NetworkX graphs for traversal. Additionally, the integration of a machine learning model from Scikit-learn, a memory-efficient library, ensures effective prioritization of edges without excessive computational overhead. In this project, we address a key limitation of the traditional Breadth-First Search (BFS) algorithm: its inability to consider edge weights during traversal. It is unsuitable for scenarios where varying edge weights significantly impact the traversal outcome, such as in shortest-path calculations for weighted graphs. To overcome this drawback, our project integrates a machine learning (ML) model to analyze and prioritize edges based on their weights, effectively augmenting BFS for weighted graphs. By leveraging CSV-based storage and combining it with an ML-driven edge prioritization mechanism. this project offers a scalable solution for managing and analyzing large, weighted graphs. This combination ensures that the navigation system not only computes the shortest path but also suggests the most practical and efficient routes tailored to user preferences or constraints.
Authors - Dipak Ligade, Chiranjit Das, Rupali Parte, Masira Kulkarni, Shivraj Jadhav, Abhishek Mohite Abstract - —This project aims to automate tree counting and forest land diversion assessment through satellite image combined with advanced computing techniques. The treatment of forest re sources needs accurate monitoring because growing environmental challenges such as deforestation, biodiversity loss, and climate change require it for sustainable land management. The research uses satellite imagery along with machine learning and deep learning tools, specifically convolutional neural networks (CNNs), to precisely detect and count trees across expansive territories. The study demonstrates how satellite analytics technologies will enhance forestry applications with their capabilities for better tree enumeration at higher efficiency and greater accuracy.
Authors - Suchanta Ravan, Prashant Dhotre Abstract - Conventional identification techniques that depend on privacy concerns and credentials are becoming more vulnerable to web-based risks like hacking and data breaches. The need for sophisticated authentication techniques has grown dramatically because of identity theft, cyberthreats, and illegal access. Traditional security methods, such as PINs and passwords, are insufficient for high-security applications since they are vulnerable to phishing, brute-force assaults, and credential breaches. To improve safety and tackle problems like privacy threats, spoofing, and accessibility problems, this study suggests a strong adaptive authentication mechanism that integrates biometric along with behavioral assessment. The multimodal authentication framework guarantees a smooth and easy verification process while also enhancing security. This structure guarantees a smooth and safe authenticating process by utilizing cutting-edge security methods like encryption, machine learning, and multifaceted biometrics in conjunction with a user-centric architecture. By combining behavioral biometrics with conventional authentication techniques, total authentication reliability is increased, and cyber risk is mitigated. The effectiveness of the suggested approach in lowering susceptibility to cyberattacks while preserving superior usability and consumer satisfaction is demonstrated by experimental findings, Highlighting the importance of two-way authentication.
Authors - S M Boomika, C M Tulasi, Sharvani V Nagur, Bhagyashri Badakali, Nalini C Iyer, Preeti Pillai, Ujwala Patil Abstract - LiDAR and cameras play a vital role in autonomous vehicles by providing complementary data for object detection and environmental perception. However, achieving seamless data integration from these sensors depends on partial and temporal synchronization. Unlike conventional methods that depend on pre-calibrated datasets, our methodology utilizes a custom-acquired multimodal dataset comprising both image and video data from a monocular camera and point cloud data from a VLP-16 Velodyne LiDAR sensor. In this paper, we proposed a comprehensive framework for LiDAR and camera calibration and temporal synchronization of real time data, synthesized and validated in a controlled lab environment. Calibration of the raw data was performed using a checkerboard as the target to ensure accurate spatial alignment between heterogeneous sensor systems.The collected corpus is further timestamped, synchronized, and validated.The accuracy of the proposed methodology is evaluated by projecting LiDAR points onto image frames, enabling qualitative verification of spatial and temporal consistency. The proposed method integrates target-based calibration with software-level timestamp synchronization to create a reproducible and scalable calibration pipeline. Results demonstrate accurate alignment across modalities, validating the effectiveness of our approach. This 1 work provides a practical contribution to multi-sensor fusion research, especially for applications requiring custom datasets or operating in constrained environments.
Authors - Wendrila Biswas, Arunangshu Giri, Dipanwita Chakrabarty, Dibyendu Rath Abstract - The study has examined the effect of user engagement (UE), perceived benefit (PB), and perceived risk (PR) of wearable sensor-based healthcare devices adoption. User empowerment (UEM) in IOT-enabled healthcare has been explored on the basis of two established theories, Technology Acceptance Model (TAM) and Behavioral Reasoning Theory (BRT). A cross-sectional online survey was conducted from November 2024 to January 2025 involving 361 valid Indian respondents and the collected responses were analyzed through NVivo software for qualitative analysis. SEM (structural equation modeling) was done for quantitative analysis and hypothesis testing. The findings have shown a positive association between UE and PB and between UE and PR. Again, the study has revealed that PB and PR positively influenced UEM. The study contributes both to existing literatures and making managerial decisions by establishing how benefits from wearable sensor-based healthcare devices can be explored by avoiding the perceived risk of the consumers and how they can get empowered with the same.
Authors - Vaishali Langote, Siddhesh Kulkarni, Aaditya Ghorpade, Aditya Songirkar, Aditya Chincholkar Abstract - Identifying customer retention is essential for decreasing lost revenues as well as maintaining an established base of loyal customers. By reviewing historical data that includes customer demographics, purchasing habits and behaviours, businesses will be able to determine which customers are going to discontinue using their services or products. In generating models that can identify customers at risk, this process includes machine learning models such as decision trees, logistic regression and neural networks. It is important that predictive retention can work provided the right algorithms are selected, and reliable data is sourced. Continual updates and improved models will enhance accuracy, giving firms the opportunity to keep up with changes in how consumers behave. The models will also give businesses the ability to produce more targeted retention marketing plans since they will not only identify at-risk customers but also give clear data on what they are doing to create customer churn.
Authors - Sharon Koshy, Padmadas Sundaram Abstract - The intensifying depletion of natural resources, fueled by world population growth and unsustainable consumption, poses severe threats to global sustainability. Specifically, the ICT and smart infrastructure industries make substantial contributions to resource inefficiencies through growing e-waste, inefficient material recovery, and unsustainable construction methods. Forecasts suggest that by 2050, with a projected 9.8 billion world population, resource use will surpass planetary limits, urging rapid interventions in resource management and the transition to circular economies. Despite growing recognition, inefficiencies in recycling infrastructure, defective waste-to-energy technologies, and inadequate water management persist to drive global resource insecurity and further environmental degradation. Solutions must be backed by evidence-based policy design, technological development, and systemic change. In this context, the combination of Artificial Intelligence and biomimicry offers a new way to increase sustainability and resilience in systems. AI-based models improve resource efficiency, reduce environmental footprint, optimize waste management, facilitate predictive maintenance, and enhance material recovery, while biomimicry offers nature-inspired solutions for sustainable design, energy efficiency, and waste reduction. These technologies not only foster resource recovery but also set the stage for the creation of wiser, more sustainable industries and cities. In conclusion, this study high- lights the revolutionary power of ICT that AI and biomimicry make possible to create closed-loop, self-sustaining models that boost urban resilience, sustainability, and efficiency, maximize recovery of resources, minimize waste, and maximize value for a truly circular future.
Authors - Priya Surana, Sushma Vispute, Madhura Kalbhor, Shubhangi Vairagar, Pragati Ugale, Imtiyaz Syeda, Mahek Yakumsha, Ashish Suryawanshi Abstract - This research presents a YouTube Comments Analyzer that leverages machine learning and deep learning algorithms to examine and classify user comments. A large volume of comments is processed by the system, enabling it to detect key patterns, including sentiment classification and emotion detection. Using natural language processing and machine learning techniques, the tool provides meaningful insights to content creators for understanding their audience and to moderators for identifying problematic content. Researchers can also benefit by studying online commentary at scale. Our team collected video comments from various genres to train and develop the models, followed by evaluation using multiple performance metrics. The analysis tool achieves 96% accuracy in sentiment detection and 90% accuracy in emotion detection, successfully identifying complex patterns that manual evaluation often misses. To demonstrate the practical applicability of our models, we further developed a web-based application that integrates the analysis pipeline, providing an accessible platform for real-time comment analysis. This research highlights the effectiveness of automated text analysis in social media environments and demonstrates real-world applications for YouTube content management and audience engagement strategies.
Authors - Omkar Kalantre, Jyoti Joglekar Abstract - Optimized Bounding box fitting around an object is necessary for accurate localization of the Region of Interest (ROI), so that features extracted from the ROI are useful for many computer vision applications. Current methods tend to be inefficient, imprecise, and with high computational complexity. In this work a novel algorithm is presented that is designed for fitting a bounding box around an object that covers maximum part of the object as ROI,. The improvement in inserting bounding box enhances the process of recognizing, tracking, and classifying objects, which is highly valuable for applications such as surveillance, autonomous driving, and security. In this work we are proposing a novel algorithm for fitting a bounding box around an object to maximize the object area covering and for minimizing the background clutter as a part of ROI.
Authors - Mohan Sellappa Gounder, Rohan Mahantesh Kamatgi, Sharath Prabhu T M, Sanya Gupta, Seema Abstract - This research investigates the application of the DINO (Distillation with No Labels) framework, a self-supervised learning approach, for efficient road and pothole segmentation. By integrating a DINO-enhanced ResNet-50 backbone with a U-Net model, this study addresses segmentation challenges in dynamic environments. The framework employs momentum encoders, multi-crop training, and stability mechanisms to facilitate robust feature extraction without requiring labeled datasets. Through strategic fine-tuning, the model achieves precise segmentation of road surfaces and potholes, making it a promising approach for real-world applications in autonomous systems and infrastructure assessment. This study further discusses model evaluation, comparison with state-of-the-art approaches, and its implications for transportation infrastructure.
Authors - Gopal D. Upadhye, Ranjana Jadhav, Aryan Pungale, Ashish Shadija, Nikita Rajput, Pranav Pendse Abstract - A data-informed system is described for generating crop recommendations and crop yield forecast based on a variety of data sources of farmer-level soil characteristics, historical crop yield records, and meteorological variable data. In the proposed system, crop recommendations based on a classification algorithm and crop yield estimates based on a regression algorithm are provided to farmers. The data-driven crop recommendations and crop yield forecasts will improve decision-making by providing the farmer with data-based recommendations providing the productivity isolation. The data-informed system will utilize machine learning algorithms to process the data and analyze the complex interaction of the various farming agri-parameters in the farm operation. Composition of soil nutrient values, weather patterns, and historical productivity variable data will be a key ingredient in the model to provide farmers with singularly specific crop selections. Ability to yield prediction gives farmers anticipate yield of the crops, improve resource planning. The validation tests demonstrate better accuracy than traditional heuristics, improving farmer overall risk reliability and increasing efficiency, sustainability. The results shows us that the transformative role of machine learning in agriculture and the associated movement toward precision farming practices
Authors - Vanshika R Kavi, Sujata Kotabagi Abstract - Semiconductor production demands high-quality control to detect faulty wafers early on in the production process. Manual inspection and rule-based systems are conventional methods that are time consuming and error-prone. This research investigates machine learning (ML) based wafer detection on a dataset of 590 sensor readings per wafer, with wafers being labeled as good (+1) or faulty (-1). Several traditional ML models, such as Logistic Regression (LR), Support Vector Machines (SVM), K-Nearest Neighbors (KNN), and Random Forest, are tested for defect classification effectiveness. The processing of data includes handling missing values by dropping features with high missing data and using median imputation. Feature selection is done through SHAP (Shapely Additive Explanations) analysis and correlation filtering to select only the most important sensor readings. Feature scaling is done to maintain consistency in data distribution. For handling the class imbalance in the dataset, SMOTE (Synthetic Minority Over-sampling Technique) is employed to create synthetic samples for the minority class to enhance model learning. Once trained, the models are evaluated on the basis of accuracy, precision, recall, F1-score, confusion matrix, and SHAP-based explainability analysis. SVM and Random Forest perform better compared to other models with 97-99% accuracy, and KNN does not perform well because of high dimensionality. The research showcases how ML is able to automate defect detection, increase production efficiency, and minimize human inspection errors. Work for the future encompasses ensemble learning optimization, real-time deployment, and semi-supervised learning optimization for enhanced defect classification in the semiconductor industry.
Authors - Piyusha S. Shetgar, Asha V. Thalange, Rohini R. Mergu, Aishwarya Khobare Abstract - Throughout the world, the number of educational institutions has significantly increased in recent decades. But the majority of recently established universities continue to manage their resources, including their hostels, using traditional methods. These conventional methods are frequently hindered by innate restrictions that negatively impact the organization's overall effectiveness. This study suggests an automated hostel lodging management system that is made with Microsoft Access as the underlying database and Visual Basic as the programming language to handle these issues. To stop unwanted access, the system has an integrated authentication algorithm. The system that has been built leverages face recognition technology to address the shortcomings of conventional approaches. It provides a graphical user interface, dependability, efficiency, and improved security by implementing access control mechanisms.
Wednesday August 26, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Associate Professor & Head, Department of Computer Science & Engineering, CSPIT, Charotar University of Science & Technology (CHARUSAT), Gujarat, India
Thursday August 27, 2026 9:28am - 9:30am IST Virtual Room DGOA, India
Authors - Ajit Patil, Amol Potgantwar Abstract - In the era of Industry 4.0, accurate time series prediction is crucial for extracting valuable insights from high-frequency sensor data in Industrial Internet of Things (IIoT) applications. This paper presents EERA: A Hybrid Ensemble Regression Model designed to improve predictive accuracy for time series data in IIoT environments. EERA combines the strengths of multiple base models, including REPTree, SMOreg, and Multi-Layer Perceptron (MLP), through a weighted ensemble approach to achieve better overall performance. The model was tested using a real-world dataset that captures heat index data (temperature and humidity), which has diverse applications in areas such as agriculture, weather forecasting, and enterprise maintenance. Comparative analysis shows that EERA outperforms individual models, achieving a Mean Squared Error (MSE) of 4.150960 & R-squared value of 0.872540, demonstrating high predictive accuracy. These findings suggest that EERA is a dependable &1 effective solution for time series prediction in fast-paced IIoT data environments.
Thursday August 27, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Manikrao Dhore, Parth Mahajan, Pratik Meshram, Ashish Nikam, Samarth Otari Abstract - Deforestation and improper plantation of trees are the key issues in realizing sustainable environmental management. The AutoForest PlantBot, an autonomous robot system, is introduced in this paper, which makes use of advanced image processing, path optimization, and real-time navigation for efficient tree plantation. The system employs the Deep Forest Package for 92% accurate tree detection and uses Dijkstra's algorithm to find optimal routes, cutting tree removal by 40% as compared to traditional straight-path approaches. The hardware system consists of an Arduino-controlled rover with BO motors, a GPS module, ultrasonic sensors, and an automated drill mechanism, providing accurate plantation with an accuracy of ±2 cm. The outcomes validate the system's potential for large-scale reforestation applications. Future developments will emphasize integrating reinforcement learning for adaptive path optimization and using renewable energy sources for sustainable operation.
Authors - S. Rahul, Anusha Preetham, Aniketh Patil, Abhishek Nimbal, Sahana Meti Abstract - This paper introduces Dr. BOT, a comprehensive web-based healthcare application designed to overcome language barriers in medical communication across diverse linguistic environments. While initially trained to predict several diseases including diabetes, heart disease, kidney disease, liver disease, and breast cancer, the system's architecture enables expansion to detect and interpret a wide range of medical conditions. Dr. BOT employs robust machine learning algorithms (Random Forest, Support Vector Machine, and Logistic Regression) trained on validated datasets, with special emphasis on multilingual functionality through a hybrid approach combining NLP with neural machine translation models specifically finetuned for medical terminology. The platform operates effectively in low-connectivity environments through innovative offline capabilities, offering preventive healthcare guidance and localized medical resource information in users' native languages, thereby supporting both individuals and healthcare providers in improving health outcomes globally.
Authors - Prasad Chaudhari, Ritesh V. Patil, Parikshit N. Mahalle Abstract - The Agricultural Productivity Enhancement System leverages data-driven pattern classification and machine learning-based fertility detection to improve farming efficiency. The architecture integrates IoT sensors, satellite imagery, and soil analysis to collect crucial agricultural data. A preprocessing module ensures data cleaning and feature extraction, storing refined data in an agricultural repository for further analysis. Machine learning models, including pattern classification and fertility detection, process this data to assess crop health and soil fertility. A decision support system then provides real-time recommendations to farmers, enhancing precision agriculture. Researchers and data analysts contribute to model refinement, ensuring scalability and adaptability. This system optimizes resource allocation, reduces wastage, and increases crop yield by enabling real-time, AI-driven decision-making.
Authors - Shiva Jyoti, Samriddhi Ganguly, B Sri Soumya, Nachiyappan S Abstract - This paper presents a comprehensive study on the Clinical Readiness Score (CRS), a structured evaluation metric for assessing AI models used in lung cancer diagnosis. The CRS incorporates multiple criteria such as interpretability, efficiency, clinical validation, and accuracy, employ- ing the Analytic Hierarchy Process (AHP) for weight assignments. This study discusses the methodology behind CRS, validates its consistency, and explores its practical implications. Additionally, graphical represen- tations of AHP weight distribution, sensitivity analysis, and CRS factor contributions are provided for better comprehension.
Authors - Harjas Singh Bajwa, Lokesh Jayakar, Abhiyanshu Singh, Prafulla Bafna, Mukta Deshpande Abstract - Builders often open sales of their property in India as soon as they purchase out land, they do this in order to secure funds to carry out their construction operations, for this they often make rendered photos and videos of concept property. Traditional property marketing techniques like images and videos lack interactivity and fail to provide a 360-degree view of properties. This research work explores the role of metaverse-driven, gamified, interactive property tours in enhancing pre-construction sales. Unlike previous studies focusing on metaverse real estate as an investment platform, this research emphasizes its ability to engage buyers, boost confidence, and aid decision-making. By combining insights from virtual real estate, Augmented Reality (AR) , virtual Reality VR, gamification, and Artificial intelligence (AI) customization, a metaverse-based property visualization framework is proposed. The study highlights how interactive walkthroughs, real-time customization, and immersive storytelling increase trust and engagement. Gamification elements, such as virtual staging, achievement systems, and AI-led personalization, deepen buyers’ connection with properties.
Authors - Ajay V, Sharon P S, Philomina Simon, Ambily George, Mehanas Shahul Abstract - This work delineates an inquiry into how artificial intelligence identifies multiple emotions in texts. Unlike mere sentiment analysis, which is a simple positive, negative, or neutral classification of text, multilabel emotion classification requires a more intense understanding of the text. The paper examines various challenges in multi-label emotion classification, where emotions often overlap (e.g., joy and surprise) and have varying frequencies in datasets. Traditional machine learning and deep learning based models such as BERT and other transformer-based models, show sufficiently strong performance in capturing nuances of emotional expression in text. Moreover, it addresses the issue of how this task can be distorted by linguistic and contextual diversity and diversity and therefore how such systems should be evaluated with respect to these variables.
Thursday August 27, 2026 9:30am - 11:30am IST Virtual Room DGOA, India
Authors - Gowri Shaju, Lekha S Nair Abstract - Histopathology refers to the study of a disease at a cellular level which stands as a golden method of predicting breast cancer. In this paper, a comparative study of the performance of machine learning models trained using optimized feature sets is done. The experiments are conducted using two datasets. The first is the Wisconsin Breast Cancer dataset, which contains 30 extracted features of cell nuclei. The second is the MITOS-ATYPIA 14 dataset, consisting of histopathology images, from which hand-crafted features have been extracted. Population based metaheuristic optimization algorithms are used to optimize and choose the key features from the available feature set to increase the efficacy of the model. Support vector machines, logistic regression model and other classification models are tested using this optimized feature set. To evaluate the impact of feature optimization, accuracy, precision, recall, and F1 score are assessed using both the full feature set and the optimized subset from two datasets. The results demonstrate how model performance varies with different feature sets, underscoring the significance of optimization techniques in enhancing machine learning-based breast cancer diagnosis in medical imaging.
Authors - Kalyanasundaram V, Keerthi AJ, Krishnaa RK, Thirumurugan A, Joshua Sunder David Reddipogu Abstract - The volatility of the stock market offers a big challenge to traders depending on timely and correct information for informed decision-making. Security risks, including fraudulent practices and theft of identity, threaten online trading platforms. The present paper presents an AI-based stock trading app that tackles these issues by using predictive analytics with robust security features. The system also employs Azure AutoML to work through historical stock data, identify market trends, and generate livestock predictions to enable traders to respond proactively to fluctuations. For security reasons, the app employs Azure Document Intelligence for live Know Your Customer (KYC) verification to ensure that only valid users have access. Additionally, the platform automates document processing using AI-powered text extraction, minimizing errors from manual input and increasing efficiency. Developed with Flutter for smooth cross-platform use and backed by Azure cloud infrastructure for scalability and dependability, this software solution offers an intelligent, secure, and user-friendly trading experience. Through the integration of AI-based forecasting with robust security measures, this work helps develop more efficient, reliable, and technologically sophisticated stock trading platforms.
Authors - George Sebastian, K. Ananda Krishnan Menon, Migheal Newton, Sonal Shaju, Haneesh K. M Abstract - Deep Vein Thrombosis (DVT) is the formation of blood clots in the lower limb because of prolonged immobility. Such critical medical conditions can be avoided by regularly using compression cuffs on the limbs; however, traditional compression devices lack adaptability, cause patient discomfort, and have inconsistent pressure application. This study presents a smart wearable and portable compression system integrated with sensors to receive real-time feedback. A DC motor, controlled by an H-bridge converter, inflates the system’s inflatable sleeves. The DC motor speed controls the pumping pressure, and a solenoid valve controls the inflation rate. Pressure, temperature, and moisture sensors are embedded in the inner part of the cuff to monitor the physiological parameters. An Arduino-based control system was used to control the inflation rate, air pressure, and duration of compression, optimally ensuring patient comfort. The designed pump was tested and shown adaptability when the sensor data changes. An AI-based control framework is also proposed in this work to enhance the performance and to make the pump autonomous and user-friendly. The response of the proposed AI-based control was validated through simulations of the model developed from fundamentals. The simulation results suggest that the AI-based DVT pump is more adaptable to the physiological parameter variations, even when the parameters change rapidly. The AI-driven model provides faster and more precise control of inflation and deflation patterns, preventing overheating, over-compression, and sweating. This study highlights the feasibility of a smart, wearable DVT pump that can adapt to the compression requirements while ensuring safety and comfort.
Associate Professor & Head, Department of Computer Science & Engineering, CSPIT, Charotar University of Science & Technology (CHARUSAT), Gujarat, India
Thursday August 27, 2026 11:30am - 11:32am IST Virtual Room DGOA, India
Authors - Gunjan Tewari, Divyanshi Verma, Simran Negi, Richa Jain Abstract - Cloud computing has completely revolutionized the way of data storage and management in a scalable environment with cost efficiency and scalability. However, the shift from physical to cloud infrastructure raises significant concerns related to security. The major security issues in cloud storage are: data breaches, unauthorized access, hacking of data, and insider threats, various types of cyber-attacks, etc. Many advancements have been made in cloud security but despite that several challenges still exist. Many security frameworks rely on third-party providers creating potential risks of data exposure. This paper addresses these challenges by proposing an approach for secure file storage in the cloud having multiple layers of security. The key derivation for encryption is done uniquely and then AES-256 is used for encryption and decryption. At the time of decryption, OTP authentication is done using RSA signing which provides multi-factor authentication. The method can also be used to encrypt all multimedia data. To provide security, the file password and OTP information are not stored in any database. The proposed work provides a very secure data storage solution that protects the data from any kind of brute-force attacks and other cryptanalysis attacks.
Authors - Jalindar Nivrutti Ekatpure, Dinesh Bhagwan Hanchate Abstract - This review paper gives a thorough look at all the current methods and uses of AI in crop prediction with aerial images. With the development of drone technology and high-resolution satellite imagery, gathering data on farming has been easier. This paper completely analyses the use full uses of Artificial Intelligence techniques in agricultural functions. The real-word ex-ample shows that how artificial intelligence techniques used in aerial imagery technology it may be accurately applied in different agricultural fields. These examples shows capability to develop observing, expect yields, and assist farmers with correct decisions. The next research enterprises have been suggested to address current difficulties and increasing artificial intelligence application in the crop cultivation techniques. This paper aims to train farmers experts, educators, and those who are to know how to use artificial intelligence in the precision farming.
Authors - Rohit Bajirao Khedkar, Bhakti Dudile, Laksh Rupesh Khobragade, Pawan Babanrao Avhad, Santosh Kumar Abstract - This paper presents a comprehensive study on the development and implementation of a virtual mouse system using hand gestures. With the rapid advancement of human-computer interaction (HCI) technologies, touchless interfaces have gained immense popularity. The proposed system eliminates the need for physical input devices by leveraging computer vision and machine learning techniques to interpret hand movements, translating them into cursor control and command execution. The research explores various methodologies, including hand tracking, gesture recognition, system integration, and deployment strategies, highlighting advancements, challenges, and future directions in the field. The objective of this study is to design an intuitive user interface, develop a robust gesture recognition model, and ensure seamless deployment across various platforms to enhance accessibility and usability.
Authors - Mariela Todorova, Tihomir Dovramadjiev, Darina Dobreva, Tsena Murzova, Mariana Murzova, Iliya Iliev, Ventsislav Markov Abstract - The pursuit of enhancing the quality and precision of final design models has become increasingly vital in modern production processes, particularly in the realm of university information signage. This research explores the application of advanced laser engraving and cutting technologies to achieve maximum accuracy in geometric shapes, fine details, and textual elements. A systematic methodology has been developed and implemented, focusing on the production of custom-designed metal information signs tailored to university environments. The article presents a comprehensive overview of the production stages, including optimized workflows for managing digital data, selection of appropriate file formats, and precise laser machine settings. By integrating digital design tools with laser technology, the study demonstrates how to streamline processes while maintaining exceptional quality and durability of signage. The research outcomes not only emphasize the technological advantages of laser systems—such as high-speed production, cost efficiency, and unparalleled precision— but also highlight their potential to transform design practices in educational settings. This study aims to contribute to the scientific and practical development of digital fabrication methods, inspiring wider adoption of laser-based innovations across design disciplines.
Authors - Devashish Sanjay Gaikwad, Aadit Kisanrao Palande, Prasad Padmakar Joshi, Aditya Atul Kode, Rachana Yogesh Patil Abstract - Assessing the solar potential of rooftops is crucial for optimizing photovoltaic (PV) installations and promoting renewable energy adoption. This study presents a methodology for estimating rooftop solar potential using advanced geospatial and machine learning techniques. The proposed framework integrates Mapbox GL for spatial visualization, PVGIS for solar radiation data, and Scikit-Learn for predictive modeling. A web-based application is developed using React.js, HTML, and TailwindCSS for the frontend, with Node.js and Express.js handling backend processes. The system allows users to input rooftop data, analyze solar potential, and generate estimations of energy output based on historical and real-time solar radiation data. By leveraging machine learning algorithms, the model enhances prediction accuracy and enables better decision-making for solar energy investments. The results demonstrate the feasibility and effectiveness of this approach in providing precise and user-friendly solar potential assessments. This research contributes to the growing field of smart energy solutions and supports the transition to sustainable energy sources.
Authors - Shaveta Thakral, JyotiVerma, Suchita Ganage, Dharmendra Ganage, Monika, Shankar Amalraj Abstract - Cognitive Radio (CR) is a revolutionary technology aimed at optimizing the utilization of the electromagnetic spectrum, a limited and valuable resource. Despite its promising potential, the deployment and widespread adoption of CR face several technical, regulatory, and practical challenges. This paper presents a comprehensive study of the key research challenges in Cognitive Radio Networks (CRNs). We begin with a chronological literature survey, highlighting significant advancements and ongoing research efforts. Subsequently, we delve into the current research challenges, including spectrum sensing, dynamic spectrum access, security, energy efficiency, interoperability, and regulatory issues. We also explore potential research opportunities that could address these challenges, thereby paving the way for more robust and efficient CRNs. This review aims to serve as a foundational reference for researchers and practitioners in the field, offering insights into future research directions.
Authors - Sheela Chinchmalatpure, Atharva Bondarde, Atharva Joshi, Archit Bagad, Samyak Dawle, Rajeshwar Chintawar Abstract - Proper waste management is essential for public health and urban sanitation. Conventional garbage collection systems tend to be absent of verification checks to confirm the emptying of waste bins and instead depend on manual records. To improve monitoring of waste collection, this research suggests a smart, technology-based solution that combines Near Field Communication (NFC) with computer vision through a TensorFlow Lite-based YOLO model. The system includes a mobile app that scans NFC tags on trash bins, logging the time and staff member who serviced the bin. The app also includes a lightweight YOLO model in TensorFlow Lite format to check if a bin is full or not using real-time image processing. NFC scanning is only allowed after successful model verification to avoid fraudulent reporting and ensure accountability of sanitation workers. This two-in-one system serves as both a real-time waste collection monitoring system and an automated worker attendance tracking system. Evaluated on a bin image dataset, the solution showed encouraging accuracy in empty and full bin detection. Through the use of AI and IoT-based tracking, this system promotes accountability, transparency, and effectiveness in waste collection, and makes it a scalable and affordable model for smart cities.
Thursday August 27, 2026 12:30pm - 2:30pm IST Virtual Room DGOA, India
Authors - Vishal V. Mahale, Sanket R. Malode, Sudarshan M. Pagare, Punit Chaudhari Abstract - Personality prediction plays a key role in understanding human behavior, decision-making, and social interactions. The OCEAN model—comprising Openness, Conscientiousness, Extraversion, Agreeableness, and Neuroticism is widely used for assessing personality traits. With the rise of machine learning, predicting personality using this model has become a growing interdisciplinary field. This survey paper reviews existing machine learning approaches, such as K-Means and Gaussian Mixture Models, used to analyze personality traits from questionnaire data. It also highlights the limitations of past studies, including lower prediction accuracy and challenges in model interpretation. The aim is to provide a clear overview of current methods and explore how machine learning can improve personality prediction and reveal deeper links between personality traits and behavior.
Authors - Poonkuzhali S, Shreya Sai Prabakar, Sriram Venkat P Abstract - The airline industry is characterized by fluctuating demand, intricate pricing, and inventory management. In that regard, this study is done regarding AI-driven revenue management systems, where many authors engage quite heavily with techniques of demand forecasting, pricing optimization, and inventory control. With Random Forest Regression, Catboost and LightGBM passenger demand can be predicted using fare class, lead time, and seasonality, optimizes seat allocation by fare category so as to maximize revenue. Evaluation is done using several datasets regarding booking patterns and market behavior in order to critically assess the accuracy, efficiency, and adaptability of the model. This study will demonstrate the strengths and trade-offs of AI techniques, indicating to the airline the power of using data for real- time decisions. Its strength lies in the improvement of demand forecasting combined with dynamic pricing and inventory management, maximized profitability, efficiency, and customer satisfaction.
Authors - Vijayalakshmi B, Jayasheela C S Abstract - Postmenopausal women (PW) are at a significantly increased risk of fractures, largely due to estrogen deficiency leading to osteoporosis and altered bone quality. Fracture risk assessment is critical for early intervention and prevention strategies. This review explores advancements in fracture risk evaluation, highlighting experimental methodologies, clinical applications and emerging technologies. Key advanced imaging approaches include the use of dual-energy X-ray absorptiometry (DEXA) for bone mineral density (BMD) measurement, high-resolution peripheral quantitative computed tomography (HR-pQCT) for micro-architectural assessment and biochemical markers like C-terminal telopeptide (CTX) and procollagen type I N-terminal propeptide (PINP) for monitoring bone mass density. Fracture risk assessment (FRA) tools such as FRAX and the Garvan calculator provide practical frameworks for estimating fracture probability by integrating clinical risk factors and BMD data. However, challenges remain. including limited access to advanced imaging, variability in biochemical marker and under representation of diverse populations in validation studies. Future directions emphasize integrating artificial intelligence, expanding population specific validations and combining imaging with dynamic bone mass density data. This comprehensive review emphasizes the significance of a multidisciplinary approach in FRA, aiming to enhance precision, accessibility and clinical outcomes in postmenopausal women.
Authors - Dhruv Aswani, Aman Sande, Praful Pradhan, Rajveer Tolani, Pallavi Saindane Abstract - Security concerns and the empowerment of women remain highly pressing challenges in India, with significant issues evident across urban, semi-urban, and rural regions alike. Women’s mobility is often constrained by the fear of harassment, crime, and social barriers which slows down their movement toward true empowerment. To address these problems, this study focuses on the major drivers of women’s safety and empowerment which include social norms, presence of crime, supportive structures, and women participation in technology. To address these challenges, this paper proposes Agati, an Android application that offers safety and empowerment features specifically tailored for women. Agati combines real-time safety alerts, location tracking, community support networks, and financial literacy modules to foster both security and economic independence. By leveraging data analytics and user feedback, the app aims to build a personalized, data-driven solution that bridges the gap between security and empowerment. Through this integrated approach, Agati seeks to create a safe, supportive environment that promotes social power and holistic growth for women.
Authors - Nikhil Vaishya, Amey Sawant, Mayank Shukla, Suhani Pandey, Vaishali Kosamkar Abstract - Augmented Reality (AR) is transforming the learning experience in anatomy and biology [1, 2]. by providing an engaging and interactive alternative to traditional teaching methods. Understanding complex anatomical structures has historically been challenging due to the limitations of textbooks, static models. AR overcomes these challenges by enabling students to explore high-fidelity 3D representations of the human body in real-time, fostering deeper spatial understanding and retention. The technology allows learners to interact with anatomical structures, receive immediate feedback, and learn at their own pace, beyond the constraints of the classroom. In addition to AR visualization, this project integrates an AI-assisted quiz and learning platform to further enhance anatomy education. By leveraging machine learning algorithms such as ensemble methods like Random Forests and Support Vector Machines (SVMs), coupled with SMOTE for class imbalance handling and cross-validation for robust generalization, the system offers adaptive quizzes, personalized learning recommendations, and real-time feedback. The platform dynamically adjusts to user interactions, ensuring a tailored and effective learning experience. Developed using Unity for AR functionalities, JSON for data management, and machine learning for prediction models, the system bridges the gap between theory and practice while promoting active and self-paced learning. This paper details the design, development, and evaluation of the AR-based Anatomy Learning Platform, highlighting its potential to revolutionize anatomy education by offering an accessible, immersive, and personalized approach. The platform is designed primarily for medical students, but it also supports general learners seeking to enhance their anatomical knowledge through immersive technologies.
Authors - Shital Pawar, Parag Dolhare, Saish Fatangare, Harshdeep Gawhale, Aditya Gadgil Abstract - Environmental, social and governance (ESG) criteria have become essential to assess the sustainability and social impact of companies. This article presents the development of an automated ESG ranking system that uses natural language processing (NLP), sentiment analysis, and machine learning techniques to rank and rate companies based on ESG metrics. Using a pre-existing database of news articles, we used the VADER sentiment analysis tool to assess the polarity of the text data, categorizing it as positive, negative or neutral. Sentiment scores were converted to numerical scores for each ESG component. In addition, Node2Vec is integrated to create network graphs that represent the relationships and interconnections between companies, allowing a comprehensive analysis of potential impacts. The results were visualized with Altair to provide a clear view of ESG trends and relationships that impact the company's performance over time. This study demonstrates the utility of combining NLP and advanced graph analytics for scalable data-driven ESG assessment.
Authors - Chalamalasetty Nishitha, Yelavarti Kalyan Chakravarti, V. Esther Jyothi Abstract - Sensitive domains such as healthcare institutions are increasingly relying on Federated learning for data security. Irrespective of this approach, they are gullible to adversarial attacks such as poisoning attacks and confidentiality breaches. To overcome these hindrances, Blockchain driven Federated learning is put forward, which integrates Secure Multi-Party Computation (SMPC) with Zero-Knowledge Proofs (ZKPs). This framework strives to ensure confidentiality in a distributed training environment. The individual entities train their local AI models with their exclusive datasets and generate Zero Knowledge Proofs to assert the accuracy of the model updates. The SMPC protocol encrypts the model updates, which are later aggregated to enable computing that guarantees privacy. Later, Smart Contracts are used to immutably store these adjustments on the Blockchain ledger, ensuring impenetrable model ensemble. To improve the trade-off between model dependability and precision, privacy noise is dynamically adjusted by employing adaptive differential privacy, based on individual client’s reputation. Extensive experiments prove the fact that the proposed system prominently reduces computing overhead in comparison to the established system while strengthening the attack detection rates. This architecture establishes a benchmark for information security in delicate areas like healthcare systems while designing its data-sensitive Al models.
Authors - Nandana R, Rithika Kannan, Ramgeeth N Nair Abstract - The rise of fintech applications has revolutionized financial decision-making, yet the determinants of risk-taking behavior in these digital platforms remain a critical research area. This study investigates the role of gamification, financial knowledge, and psychological influences in shaping users’ risk-taking behavior. Using a quantitative approach, an Ordinary Least Squares (OLS) regression analysis was conducted on a dataset of 200 fintech users. The results indicate that gamification has a significant positive effect on risk-taking behavior (β = 0.1414, p = 0.001), suggesting that game-like elements in fintech apps encourage users to take greater financial risks. However, certain gamification effects exhibit a negative influence (β = -0.1272, p = 0.005), highlighting that not all gamification strategies lead to in- creased risk-taking. Financial knowledge also emerged as a significant determinant (β = 0.1965, p = 0.001), implying that financially literate users tend to take more calculated risks. Among psychological factors, risk tolerance (β = 0.2754, p < 0.001) was the strongest predictor, demonstrating that individuals predisposed to risk-taking in general extend this behavior to fintech platforms. Additionally, social efficacy (β = 0.2461, p < 0.001) and social influence (β = 0.1598, p = 0.004) significantly contribute to risk-taking, emphasizing the role of self-perceived competence and peer influence in financial decision-making. The model explains approximately 48.1% of the variance in risk-taking behavior (R² = 0.481), confirming the robustness of these deter- minants. The findings underscore the importance of designing fintech applications that balance engagement with responsible financial behavior. Future research should explore the ethical implications of gamification and assess long-term user behavior to ensure sustainable financial decision-making in digital finance ecosystems.
Authors - Ramesh Babu Mutluri, Vinit Kumar Singh, D Saxena Abstract - Rural areas in developing countries are still refrained from continuous and uninterrupted power supply to power their household and run small industries. Thus, we can say that these rural areas are weakly connected to the utility grid. The main reasons for poor power supply are weak infrastructure, lack of adequate generation to fulfill the demand-supply gap, dependency on long-distance transmission, frequent load shedding, and distributed generation. This demand-supply gap can be minimized by installing renewable energy sources with the local load forming rural microgrid and connecting to the utility grid. The grid connection would help to maintain the power supply due to the variable output characteristics of renewable energy sources thus also acting as a buffer to the local power system. This paper presents a novel approach towards modeling of utility connected rural microgrid comprising renewable energy sources considering control architecture for marinating frequency-voltage interdependency. Accordingly, a frequency-based voltage controller is introduced. Further, the model has been verified in view of various scenarios with a fluctuation in load demand and power input to renewables. The controllers are tuned such that in case of increase in load or decrease in power generation, power demand is met from the utility grid, and in case of surplus generation, the power is fed to the grid, therefore, developing microgrid as business unit applicable for power trading. The model has been developed in Simulink/MATLAB. An integral square error criterion has been used for tuning the controllers to mitigate the oscillations.
Authors - Ashwitha A Shetty, Naganna Chetty, Antony P.J Abstract - The poultry industry is a significant and prominent business sector. As the daily intake of chicken meat and eggs is rising globally, poultry farming is gaining significance for providing protein. Additionally, this industry raises the nation's revenue despite being a less expensive protein source. Numerous diseases that harm the chickens are the main issue affecting the poultry business. Due to the high cost of vaccinations, poultry owners are unable to adopt these expensive methods. Consequently, this strategy cannot be used because it requires continuous investment. This paper aims to present one of the prevalent chicken diseases, coccidiosis and the different detection techniques used. In this regard, the study introduces multiple strategies that can be used in tandem to identify coccidiosis-affected fowl hens automatically. The idea behind studying chicken activity monitoring is that it directly connects to the health condition of the chicken. The enhanced future research could result in a system to monitor chicken activity and detect coccidiosis among them
Thursday August 27, 2026 3:30pm - 5:30pm IST Virtual Room DGOA, India
Authors - Piyush Sharma, Harish Patidar, Anuj Kumar Abstract - This research introduces a ResNet-based framework for multiclass classification of mammographic density and mass regions. The framework was rigorously tested using two prominent mammographic datasets, INbreast and DDSM, and benchmarked against other models, including CNNs, Random Forest (RF), Support Vector Machines (SVMs), Logistic Regression (LR), and K-Nearest Neighbors (KNN). ResNet demonstrated superior performance across all critical evaluation metrics—accuracy, precision, recall, F1-score, and AUC—outclassing the comparative models on both datasets. Its proficiency in extracting complex hierarchical features and addressing multiclass classification tasks positions it as a robust choice for breast cancer diagnosis. This framework offers a reliable and efficient tool for automating diagnostic processes, with the potential to significantly improve clinical decision-making and patient care.
Authors - Pranav Bagal, Bhavesh Patil, Shounak Muglikar, Yash Sonavane, Prajakta S. Shinde Abstract - The study addresses dental caries detection and classification using state-of-the-art deep learning architectures. We implemented and compared three pre-trained convolutional neural network models: VGG19, DenseNet169, and ResNet101, to automatically identify and classify dental caries from intraoral clinical image. Our research focused specifically on pediatric populations aged 1 to 14 years, where caries remain a significant health concern despite global prevention efforts. The models were trained and validated on a comprehensive dataset of dental images. Performance metrics demonstrated that DenseNet169 model achieved superior results with an Validation accuracy of 72.22%. These deep learning approaches show promising potential to augment traditional diagnostic methods, particularly in resource-limited settings where expert dental practitioners may be scarce. By enabling earlier and more accurate detection of carious lesions, our proposed system could help address disparities in oral healthcare accessibility and contribute to more effective intervention strategies, especially for underprivileged populations where caries prevalence continues to rise. This research establishes a technological framework that could be integrated into portable diagnostic tools for use in diverse clinical environments.
Authors - Govinda Sambare, Lalit Deore, Harsh Itkar, Onkar Jadhav, Sarthak Joshi Abstract - This research presents the development of an intelligent stock recommendation system that utilizes advanced machine learning models for informed long-term investment decisions. The system addresses the complexities of the stock market, where traditional methods often fall short in accessibility, accuracy, and efficiency. By automating fundamental analysis with models like Long Short-Term Memory (LSTM) networks and the CNN-GRU-XGBoost hybrid model, the system integrates key financial ratios, macroeconomic indicators, and sector performance, providing data-driven insights. The proposed framework optimizes stock selection using XGBoost and forecasts future stock prices with LSTM, offering precise and scalable solutions for diverse investment portfolios. The literature review highlights modern methodologies like TRAN, Bi-LSTM, and hybrid models, which improve stock forecasting and trading strategies by incorporating temporal dependencies and inter-stock relationships. The algorithmic analysis explains LSTM's ability to handle sequential data and the hybrid model's powerful feature extraction and prediction capabilities. This hybrid approach enhances decision-making, saves time, and democratizes financial insights, making advanced analysis accessible to individual investors, robo-advisors, and educational institutions. While offering benefits like scalability and reduced biases, the system also faces challenges, such as computational costs and market volatility. Backtesting results confirm the system's adaptability to dynamic market conditions, ensuring sustainable investment strategies. This project showcases the transformative potential of AI/ML in financial analytics, laying a strong foundation for long-term, informed investment decisions.
