作者投稿和查稿 主编审稿 专家审稿 编委审稿 远程编辑

计算机工程 ›› 2011, Vol. 37 ›› Issue (20): 282-284. doi: 10.3969/j.issn.1000-3428.2011.20.096

• 开发研究与设计技术 • 上一篇    下一篇

基于三维体的多分辨率视频分割算法

眭博聪,陈敏刚,高 岩,马利庄   

  1. (上海交通大学计算机科学与工程系,上海 200240)
  • 收稿日期:2011-04-19 出版日期:2011-10-20 发布日期:2011-10-20
  • 作者简介:眭博聪(1986-),男,硕士研究生,主研方向:视频图像处理;陈敏刚,博士研究生;高 岩,讲师、博士;马利庄,教授、博士、博士生导师
  • 基金资助:
    国家自然科学基金资助项目(60873136);国家“863”计划基金资助项目(2009AA01Z334);上海市科委重点科技攻关计划基金资助项目(08515810201);泰安市科技发展计划基金资助项目(20 092016)

Multi-resolution Video Segmentation Algorithm Based on 3D Body

SUI Bo-cong, CHEN Min-gang, GAO Yan, MA Li-zhuang   

  1. (Department of Computer Science and Engineering, Shanghai Jiaotong University, Shanghai 200240, China)
  • Received:2011-04-19 Online:2011-10-20 Published:2011-10-20

摘要: 传统视频分割算法无法处理视频中庞大的数据量,且分割结果需要保持时空连续性。针对该问题,提出一种基于三维体的多分辨率视频分割算法。在三维空间中对整个视频体进行分割,以保证时空连续性,采用多分辨率的思想,减少计算开销。实验结果表明,该算法能减少传统分割算法的时空耗费,修正连续帧的局部分割结果。

关键词: 视频分割, 三维多分辨率分割算法, graphcut算法, 三维环带, 局部修正

Abstract: This paper proposes a novel 3D body based multi-resolution video segmentation algorithm for solving some difficulties in video area, such as the tremendous data volume and temporally inconsistency. It segments the whole video volume to ensure the temporally coherent result. Meanwhile, it uses the multi-resolution method to reduce the computational cost. Experimental results show that, the approach greatly reduces the time and space cost of the traditional segmentation approach and revises the part segmentation result of continuation frame.

Key words: video segmentation, 3D multi-resolution segmentation algorithm, graphcut algorithm, 3D girdle, locality revision

中图分类号: