摘要： 提出一套实时视频对象分割检索算法，用运动矢量信息标注运动区域，利用背景检测算法检测对象区域。将检测出的对象区域和运动区域进行并操作运算，得到单帧图像的分割。利用改进的Hausdorff距离跟踪算法进行匹配跟踪和检索。实验表明了该算法的有效性。
关键词:
实时对象分割,
检索算法,
Hausdorff距离
Abstract: This paper proposes a real-time video object segmentation and retrieval algorithm. The algorithm marks the moving areas, and extracts the object information by using the background detection algorithm. A pixel in object areas and moving areas is set to moving object areas in a frame. Moving object is tracked and indexed by using Hausdorff distance. The algorithm is effective and real-time.
Key words:
real-time object segmentation,
retrieval algorithm,
Hausdorff distance
中图分类号:
肖 广;石旭利. 实时运动对象分割和检索算法[J]. 计算机工程, 2008, 34(7): 226-228.
XIAO Guang; SHI Xu-li. Real-time Moving Object Segmentation and Retrieval Algorithm[J]. Computer Engineering, 2008, 34(7): 226-228.