A framework for modeling appearance change in image sequences
被引:31
作者:
Black, MJ
论文数: 0引用数: 0
h-index: 0
机构:
Xerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USAXerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USA
Black, MJ
[1
]
Fleet, DJ
论文数: 0引用数: 0
h-index: 0
机构:
Xerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USAXerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USA
Fleet, DJ
[1
]
Yacoob, Y
论文数: 0引用数: 0
h-index: 0
机构:
Xerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USAXerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USA
Yacoob, Y
[1
]
机构:
[1] Xerox Corp, Palo Alto Res Ctr, Palo Alto, CA 94304 USA
来源:
SIXTH INTERNATIONAL CONFERENCE ON COMPUTER VISION
|
1998年
关键词:
D O I:
10.1109/ICCV.1998.710788
中图分类号:
TP [自动化技术、计算机技术];
学科分类号:
0812 ;
摘要:
Image ''appearance" may change over time due to a variety of causes such as I) object or camera motion; 2) generic photometric events including variations in illumination (e.g. shadows) and specular reflections; and 3) "iconic changes" which are specific to the objects being viewed and include complex occlusion events and changes in the material properties of the objects. We propose a general framework for representing and recovering these "appearance changes" in an image sequence as a "mixture" of different causes. The approach generalizes previous work on optical flow to provide a richer description of image events and more reliable estimates of image motion.