Comparison of VGG-16, VGG-19, and ResNet-101 CNN Models for the purpose of Suspicious Activity Detection

Abstract

This paper compares the performance of three popular convolutional neural network (CNN) models, VGG-16, VGG-19, and ResNet-101, for the task of suspicious activity detection. The VGG networks are known for their depth and the use of small convolutional filters, while ResNet is known for its residual connections that allow for deeper networks without the issue of vanishing gradients. The study utilizes a dataset of surveillance videos for training and testing the models. The results show that the VGG-19 model outperforms the other two in terms of accuracy, specifically for the detection of suspicious activities. Overall, this study demonstrates the effectiveness of the three different models for suspicious activity detection and highlights its potential for use in real-world surveillance systems.