Evaluating Convolutional Neural Network Models: Performance Perspective in Video Summarization

The nature of data is evolving with technological progress. Initially dominated by text datasets, the focus has now shifted to images and, more recently, to extensive video datasets. This evolution necessitates advanced technologies capable of processing images and developing intelligent systems to accurately extract information from them. Pre-trained convolutional neural network (CNN) models are essential tools for this task. In this paper, we present a comparative analysis of the performance of various CNN models, including AlexNet, GoogleNet, and SqueezeNet, specifically for image classification. We evaluate and compare the accuracy of these models in object detection across three different datasets—animals, birds, and flowers—sourced from Kaggle's online repository.