Image Classification For Visual Recommendation Using Deep Learning
Anas Laamouri, Nawal Sael · 2024
Visual recommendation systems are designed to provide users with personalized suggestions, including images, videos, products, or content, by leveraging their preferences, historical behavior, or other relevant characteristics. This versatile approach finds application in diverse contexts such as video streaming platforms, e-commerce websites, social networks, and more. This paper, conduct a comprehensive comparative study of leading Deep Learning models within the current state of the art on visual recommendation. This study places particular emphasis on the pivotal classification phase within the recommendation process. In this exploration, we compared four cutting-edge models, namely: VGG16, ResNet50V2, Xception, and InceptionV3. To evaluate these models, we used Amazon's fashion product image dataset, comprising 45 categories. The InceptionV3 model performed best, with an accuracy of 0.9395.