作者: Jin Suk Kwak , Myoung Ho Lee , Mun Chul Kim , Jae Gark Choi , Chie Teuk Ahn
DOI:
关键词:
摘要: A video object segmentation method is disclosed in the present invention. This for correctly segmenting objects. In method, objects to be firstly appeared a sequence are segmented manual or semi-manual, and then consecutive frames automatically by an object-tracking based on moving of primarily object. other words, applicable system, comprises first step existing frame manually semi-manually; second within frames. can adapted system.