An image segmentation algorithm for segmentation-based video sequence coding is presented. The method uses a new criterion based on similarities in both motion and brightness. Brightness and motion information are incorporated in a single segmentation procedure. The actual segmentation is accomplished using a region-growing technique based on the watershed algorithm. It comprises the processes of extracting the initial regions called seeds and their growing by merging similar neighboring pixels. In addition, a tracking technique is used in subsequent frames to achieve a coherent segmentation through time. Simulation results show that the proposed method is effective in deter- mining object boundaries not easily found using the static criterion alone. #1997Society of Photo-Optical Instrumentation Engineers. S0091-3286(97)01008.8 Subject terms: segmentation; homogeneity criterion; region growing; tracking. Paper 10037 received May 3, 1996; revised manuscript received Mar. 7, 1997; accepted for publication Mar. 7, 1997.
展开▼
Korea Advanced Institute of Science and Technology Department of Electrical Engineering /373-1 Kusong-dong, Yusong-gu Taejon 305-701, Korea E-mail swlee @ sdvision.kaist.ac.kr;
