SPATIOTEMPORAL VIDEO SEGMENTATION USING A JOINT SIMILARITY MEASURE

This paper presents a new morphological spatiotemporal segmentation al gorithm, The algorithm incorporates luminance and motion information s imultaneously and uses morphological tools such as morphological filte rs and watershed algorithm, The procedure toward complete segmentation consists of three steps: joint marker extraction, boundary decision, and motion-based region fusion, First, the joint marker extraction ide ntifies the presence of homogeneous regions in both motion and luminan ce, where a simple joint marker extraction technique is proposed. Seco nd, the spatio-temporal boundaries are decided by the watershed algori thm, For this purpose, a new joint similarity measure is proposed, Fin ally, an elimination of redundant regions is done using motion-based r egion fusion. By incorporating spatial and temporal information simult aneously, we can obtain visually meaningful segmentation results. Simu lation results demonstrates the efficiency of the proposed method.