Product similarity detection in E-Commerce website
Gayathri R S, S. Kanagaraj · 2023
E-commerce websites are becoming increasingly popular, and with the abundance of products available, it can be challenging for customers to find what they are looking for. One approach to address this issue is by providing product recommendations based on their similarities. In this paper, we propose a product similarity detection system using Generative Adversarial Networks (GANs). Our system generates a latent representation of the product images, which are then used to compute a similarity score between two products. We trained our GAN network on a dataset of product images, and we evaluate our system's performance on a separate test set. Our results demonstrate that our proposed approach outperforms traditional similarity measures such as cosine similarity and Euclidean distance. The proposed system can be integrated into online business sites to give clients precise and customized item proposals, ultimately improving their shopping experience.