Acta Scientiarum Naturalium Universitatis Pekinensis

Previous Articles     Next Articles

Automatic Segmentation of Moving Objects for Head-shoulder Video Sequence

LUO Tao   

  1. Courage Image Processing Lab, Mathematical School, Peking University, Beijing,100871
  • Received:1999-11-16 Online:2000-09-20 Published:2000-09-20

头肩视频图像的运动物体自动提取

罗涛   

  1. 北京大学数学科学学院,北大科瑞奇图像处理实验室,北京,100871

Abstract: Introducing and comparing the difference within the conception of area segmentation, motion segmentation and moving object segmentation (VOP extraction), and their application in content-based video sequence image coding. Analysis the relationship between the frame difference image and its relative frame gray level image. Then a new automatic moving object edge tracking and extracting method, based on the high correlation between the edge of frame difference and the edge of gray level image,was proposed. Using the close of moving object edge to get the real moving object.

Key words: segmentation, video object, area segmentation, motion segmentation, object segmentation, object tracking

摘要: 介绍、比较区域分割、运动分割和物体分割的概念及其在视频序列图像分割中的应用。分析说明各种运动状态下帧差图像与相应帧灰度图像的关系。提出了基于帧差图像边缘与灰度图像边缘之间强相关性的运动物体边界自动跟踪、提取算法。利用运动物体边界的闭包分割、提取运动物体。提出了多层次运动物体描述的思想。

关键词: 图像分割, 视频对象, 区域分割, 运动分割, 物体分割, 运动跟踪

CLC Number: