期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2015
卷号：15
期号：3
页码：36-39
出版社：International Journal of Computer Science and Network Security
摘要：Image retrieval based on visual similarity has become an active research in recent years. For each and every image we create hash code by extracting the features of the image, here search is performed based on hamming distance. Unlike Euclidean metrics that offer continuous distances, the hamming distances are separate integer values. As a result, they are often many images that share equal hamming distance to a query, where there is a chance of error occurrence so we go for ranking of image. This paper introduces an new approach where similar and class based images can be retrieved based on hamming distance. This is the first time where the concept of data mining is being introduced into image searching, which will reduce the amount of time taken to retrieve similar images.
