The Research Of Video/Image Segmentation Technology Based On JSEG

As the rapid of development of science and technology, especially the development of computer and internet, the multimedia information which includes text, image, video and audio has become the main media where people get information. As the development of multimedia information, the image/video has also been used widely. The technology of image segmentation and object segmentation of video sequence is also a very important subject, because the quality of segmentation will directly affect the subsequent work. So the research of high quality segmentation of image/video based on existed segmentation methods has great practical significance.On the aspect of color image segmentation, the paper propose an novel method based on joint JSEG and improved watershed algorithm, which can effectively reduce the computational load while basically keeping good perceptual results. Thus, the segmentation efficiency is improved. On the aspect of object segmentation of video sequence, a novel segmentation algorithm based Spatial-temporal is also proposed, in the temporal domain, we apply high-order statistics to detect the motion region, in the spatial, an improved JSEG algorithm is used to segment the motion region, finally, a moving object is extracted with projecting operation of spatial and temporal segmented results. And the results under the condition of still background show that the proposed algorithm is efficient.