Accuracy Comparison of VGG-19 Model and Combined Resnet-50 and VGG-19 Model

Abstract

VGGNet and ResNet have become popular algorithms for classifying images. This paper will compare the accuracy of the VGG‑19 model when used alone and the accuracy of VGG‑19 model when used in combination with the Resnet‑50 model. Both the models are pre trained models, and we have used these pre trained models on a dataset. The dataset consists of 133 classes of dog breeds with a total of 6680 training images and 836 test images. The objective of this study is to understand if it is more suitable to use VGGNet alone or if combining it with ResNet will provide better accuracy.