Application of CBIR in E-commerce

Abstract

The rise of technology and the rapidly increasing inventions in Science have completely changed many aspects of the world today. Many sectors such as communication, banking, media, etc. have gained momentum because of the internet. Online shopping is one such sector that has flourished in recent times because of the internet. This paper presents a method which employs the system of Content Based Image Retrieval (CBIR) in online shopping. Using this system, the time required to shop online will be reduced. CBIR is the activity of fetching images from the database which have some similarity to the given query image. Traditionally customers would have to search from different categories and apply various filters to buy the product that they want. But in this system, they will be provided with an option to directly upload the image of the product that they wish to buy. If similar products are available, it will be displayed to the customer immediately. Thus, the time required for a customer to buy a product reduces considerably thereby making the shopping experience fun, easy and convenient. The system works in a way such that when an image is uploaded, the features of this image are extracted by using the deep learning method of Convolutional Neural Network (CNN). These extracted features are compared with the features of the available images stored in the database. Then, the similarity measure is calculated and images that are akin to the query image are found and are set out as result. This method significantly helps in reducing the time required to search for a particular product.