Visual Based Image Search using Perceptual Hash Codes for Online Shopping

Abstract: Visual based search aims at identifying a product, or retrieving similar products from a database based on a photo captured from a mobile phone camera or browsed from computer. Today, Keywords are used to find products on online shopping site. The fast and easy search has motivated the development of Visual based search which is the type of search that puts images at the centre of user�s attention. Visual basedimage search works like traditional (key-word) search but without having to type any text or go through complicated menus to initiate the search process. We focus on a visual shopping system. To be more specific, we want to use visual information to find corresponding products in a precise and easy fashion. This is useful when people do not know how to describe the visual object in terms of text. In this paper we discuss the new hash technique called as perceptual hash which calculates the hash code for input image provided by user and then hamming distance is calculated with the images stored in the databases having their respective perceptual hash value and based on highest percentage of visual similarity, products are displayed to user.

Keywords: visual based image search, perceptual hash codes, hamming distance, visual similarity.