Skip to Main Content
A merged histogram method (MHM) is proposed for histogram-based image retrieval based on dominant colors in images. In MHM, colors from images and between images are merged to form their dominant color set, instead of color components. Images of similar color can thus be selected and ranked. Experimental results show that MHM can achieve up to 32% and up to 22% more precise than RGB-based histogram and perceptually weighted histogram methods, respectively.
Circuits and Systems, 2003. ISCAS '03. Proceedings of the 2003 International Symposium on (Volume:2 )
Date of Conference: 25-28 May 2003