Image retrieval based on multiple features using wavelet

Abstract

An algorithm of image retrieval using both color feature and texture feature is proposed in this paper. Circular region energy in low frequency band of wavelet transform of an image is used as color feature of the image and the synthesize energy in high frequency bands of multi-scale wavelet transforms is used as texture feature. A method using the linear combination of several single features is applied for image retrieval based multiple features, in which the weight of every feature is determined self-adaptively, moreover, relevance feedback technique is used to adjust the weights in order that the retrieval results accord with the user's retrieval goal gradually.