Midstream content access based on color visual pattern coding

It has been argued that future image coding techniques should allow for "midstream access", i.e. allow image query, retrieval, and modification to proceed on the compressed representation [18]. In a recent work [1], we introduced the colour visual pattern image coding (CVPIC) technique for colour image compression. An image is divided into blocks and each block coded locally by mapping it to one of a predefined, universal set of visually significant image patterns consisting of representations for both edge and uniform (smooth) regions. The pattern and colour information is then stored, following a colour quantisation algorithm and an entropy encoding stage. Compression ratios between 40:1 and 60:1 were achieved whilst maintaining high image quality on a variety of natural colour images. It was also shown that CVPIC could achieve comparable performance to state-of-the-art techniques such as JPEG. In CVPIC, the colour distribution information forms part of the compressed data, and hence enables direct and fast implementation of content (colour) based image retrieval in the compressed domain. Also because the edge patterns are coded separately, information about the shapes of the objects and structure orientations within an image can be easily deduced from the compressed data of CVPIC. In this paper, we shall show that CVPIC is an image coding technique that meets the "4 criterion" [18], i.e. the compressed representations of CVPIC can be directly used for content based indexing and retrieval. A method for effectively matching the compressed features of CVPIC is described and experimental results demonstrating the effectiveness of this method in retrieving colour images in the compressed domain of CVPIC are presented.