An image-based recommender system based on feature extraction techniques

Recommender systems provide recommendations about various products and services to their users while using other users data. Their success is imperative for both users and the e-commerce sites utilizing such systems. Providing accurate and dependable recommendations increases user satisfaction that results selling more products and services. Image-based recommender systems are receiving increasing attention in the recent years. These systems are based on images uploaded by users. They first determine the most similar users based on the images they upload to the system. They then return the most likely image to the users based on the images liked by the neighbors. The most challenging problem in image-based recommneder systems is to match an image with the most similar visual words or classes based on the image's visual content. We design a system which is using Bag of words model, to solving this problem.
¹
BoW model is an effective model in computer vision field, and we use SURF, SIFT and LBP descriptors for extracting features in our system.