Motion analysis and segmentation in 3D scene

被引：0

作者：

Zhang, J. ^{[1
]}

Zhu, G. ^{[1
]}

Liu, W. ^{[1
]}

Liu, D. ^{[1
]}

机构：

[1] Dept. of Electronics and Info. Eng., HUST, Wuhan 430074, China

来源：

Huazhong Ligong Daxue Xuebao/Journal Huazhong (Central China) University of Science and Technology | 2001年 / 29卷 / 08期

关键词：

D O I：

暂无

中图分类号：

学科分类号：

摘要：

An algorithm for video sequence analysis and image segmentation by using 3D scene model is presented. At the beginning of the sequence, the procedure uses two frames to acquire the depth map, and represent the scene as a 3-D wire-frame computed from the depth map. A linear algorithm with low complexity is used to recover the motion parameters, and update the 3-D scene description for each additional frame. The approach ia applied to solve the problem such as video sequence segmentation, object tracking, and video object plane (VOP) generation.

引用

页码：26 / 28