Effects of Quaghistogram on Image Retrievl and Clustering

Availability of Huge image datasets and requirement of employing them causes necessity of effective methods for managing them. Extracting efficient features from images is one of the most important requirements in managing techniques such as image retrieval, indexing, classification, clustering and etc. Identifying the descriptive and discriminatory features is valuable and essential. In this paper a color feature is proposed called QuadHistogram and we use it for image retrieval and clustering. QuadHistogram incorporates qudtree decomposition and color histogram. For extracting this feature, firstly the quadtree decomposition is applied on the images and homogenous blocks with different size are specified and then color histogram and complexity of blocks of the same size are extracted. Experimental evidence shows this feature in compare to global color histogram achieves better affect and efficiency in image retrieval and clustering.