期刊名称：International Journal of Computer Science Issues
印刷版ISSN：1694-0784
电子版ISSN：1694-0814
出版年度：2012
卷号：9
期号：2
出版社：IJCSI Press
摘要：This paper is concerned with the development of video segmentation technique for segmenting both the static and dynamic objects in the video sequence. The effective extraction of objects from the frame is an important process of video segmentation. The proposed technique uses the fuzzy C means approach for object extraction. The proposed approach uses the joint use of frame difference algorithm with the background subtraction method as well as consecutive frame difference for segmentation of static and dynamic objects. The prior research used the FKM based object extraction method whereas the proposed approach uses the improved FCM based object extraction method. The proposed technique is evaluated by varying video sequences and the efficiency is analyzed by calculating the statistical measures and kappa coefficient.
关键词：Video segmentation; Discrete cosine transform; Fuzzy C Mean Clustering; Fuzzy K Mean clustering ; backround subtraction method
