A simple scheme for motion boundary detection

被引:4
作者
Chen, WG
Nandhakumar, N
机构
[1] UNIV VIRGINIA,DEPT ELECT ENGN,CHARLOTTESVILLE,VA 22903
[2] MICROSOFT CORP,REDMOND,WA 98052
关键词
motion boundary; motion discontinuity; motion estimation; motion detection;
D O I
10.1016/0031-3203(96)00017-9
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
A new approach is developed for detecting motion boundaries and independent motion without explicit estimation of the motion field. The formulation is based on the differential optical flow equation and uses a piecewise linear model for the image intensity data. Compared with previous techniques, our approach has greater simplicity and makes less specific assumptions on the motion held. Experimental results on real image sequences and real time implementation are included. Copyright (C) 1996 Pattern Recognition Society.
引用
收藏
页码:1689 / 1701
页数:13
相关论文
共 24 条
[21]  
Wang J. Y. A., 1993, Proceedings. 1993 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (Cat. No.93CH3309-2), P361, DOI 10.1109/CVPR.1993.341105
[22]  
ZETZSCHE C, 1993, PATTERN RECOGN, V12, P771
[23]  
ZHANG J, 1993, P IEEE INT C AC SPEE, P197
[24]   An error-weighted regularization algorithm for image motion-field estimation [J].
Zheng, Heyun ;
Blostein, Steven D. .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 1993, 2 (02) :246-252