We present a constraint adaptive image segmentation techniquedesigned to achieve the combined purposes of color imagecoding/compression, indexing and content-based retrieval. An image issegmented into homogeneous squared regions of variable sizes such thatit can be encoded very efficiently. From the segmented image, we derivean effective and efficient image content description feature termed theregion and color co-occurrence matrix (RACOM) as image index forcontent-based indexing and retrieval. Experimental results are presentedto demonstrate that RACOM is a very effective image content descriptionfeature and has comparable performance to state of the art methods,colour correlogram and MPEG7 colour structure histogram, incontent-based image retrieval from large image databases
展开▼