Authors - Rachit Adhvaryu, Dipesh Kamdar, Krishna Raulji, Shreya Dholariya, Krunal Vaghela, Munindra Lunagaria Abstract - The nature of data has changed significantly as a result of the quick development of technology. Text-based datasets have given way to visual data, such as photos and videos. This shift necessitates the development of cutting-edge technologies that can effectively process and analyze visual input, allowing the creation of intelligent systems that can precisely extract insightful information. Convolutional neural network (CNN) models that have already been trained are now essential resources for this project. In this paper, the effectiveness of three well-known CNN models—ResNet, DenseNet, and VGG—in picture classification tasks is thoroughly compared. Our assessment concentrates on object detection performance on three different datasets—animals, birds, and flowers—obtained from the online repository of Kaggle, offering important information about the advantages and disadvantages of each model.
