Comparison of VGG and ResNet Performance on the Oxford Flower Dataset

Abstract

In order to compare the efficacy of vGG and ResNet architectures in image classification tasks, a thorough performance evaluation of each architecture is conducted in this report using the Oxford Flower Dataset. The main goals are to evaluate recall, accuracy, and precision while paying particular attention to the subtle distinctions between the two architectures. The research utilizes a strict methodology that includes model configuration, training parameters, and dataset preparation. The evaluation involves both training and testing phases, enabling a robust comparison. The performance characteristics of vGG and ResNet in deep image classification tasks are elucidated by the results, which also highlight their respective shortcomings. All things considered, this analysis adds important information to the current discussion about the best convolutional neural network architectures for image classification.