| 崔智高,李艾华,冯国彦.动态背景下融合运动线索和颜色信息的视频目标分割算法[J].光电子激光,2014,(8):1548~1557 |
| 动态背景下融合运动线索和颜色信息的视频目标分割算法 |
| A video object segmentation algorithm for dynamic back ground combining motion cue with color information |
| 投稿时间：2014-02-20 |
| DOI： |
| 中文关键词: 视频分割 动态背景 运动轨迹 超像素 马尔可夫随机场模型 |
| 英文关键词:video segmentation dynamic background motion trajectory superpixel Markov random field model |
| 基金项目:国家自然科学基金(61021063,61225008)资助项目 (1.第二炮兵工程大学,502教研室,陕西 西安 710025; 2.西安交通大学 机械工程学院,陕西 西安 710049) |
|
| 摘要点击次数: 1613 |
| 全文下载次数: 106 |
| 中文摘要: |
| 针对现有动态背景下目标分 割算法存在的局限性,提出了一种融合运动线索和颜色信息的视频序列目标分割算法。首先 ,设计了一种新的运 动轨迹分类方法,利用背景运动的低秩特性,结合累积确认的策略,可以获得准确的 运动轨迹分类结果； 然后,通过过分割算法获取视频序列的超像素集合,并计算超像素之间颜色信息的相似度； 最后,以超像素为节 点建立马尔可夫随机场模型,将运动轨迹分类信息以及超像素之间颜色信息统一建模在马尔 可夫随机场的能量函 数中,并通过能量函数最小化获得每个超像素的最优分类。在多组公开发布的视频 序列中进行测试与对 比,结果表明,本文方法可以准确分割出动态背景下的运动目标,并且较传统方法具有更高 的分割准确率。 |
| 英文摘要: |
| Video object segmentation is important for object tracking,video surv eillance and semantic classification.In order to overcome the limitation of existing video object segmentation methods u nder dynamic background,a video segmentation algorithm based on motion cue and color information is proposed in this paper.At first,a new motion trajectory classification method is designed.The proposed method can accurately divide the motion trajectory set into background and moving object ones by combining the low rank constraint and cumul ative acknowledgment strategy. Then,the superpixels are acquired by over-segmenting method.And the co lor similarity of adjacent superpixels is computed according to their common boundary.At last,taking the superpixels as n ode,an energy function of Markov random field model is designed,which has combined motion trajectory classificat ion information and color similarity of superpixels.The classification of each s uperpixel can be obtained by finding the minimum of the energy function. The proposed algorithm is tested on several publicly available videos.Experimen tal results demonstrate that the proposed method can accurately segment the moving objects from the dynamic backg round,and it has better segmentation accuracy compared with classical methods. |
| 查看全文 下载PDF阅读器 |
| 关闭 |
|
|
|
|
|