Garment image retrieval based on multi-features

Abstract

In order to meet the requirements of users, a content-based image retrieval algorithm for electronic commerce has been investigated combining color features with shape features. The algorithm of image retrieval using both the invariant moment and the Fourier descriptor is proposed to retrieve the shape features of image. The secondary retrieval has been done in succession by use of modified color histogram. To avoid the influence of background, a method of image background elimination is proposed prior to the image retrieval using multi-features. Experimental results show that the algorithm is accuracy. The methods of garment image retrieval based on multi-features are promising to be useful for applications of online shopping.