基于马尔可夫随机场与活动轮廓的运动目标分割  

Moving object segmentation based on Markov random field and active contour model

作　　者：仲兴荣[1] 黄贤武[1] 刘家胜[1] 

机构地区：[1]苏州大学电子信息学院,江苏苏州215021

出　　处：《计算机应用》2007年第7期1663-1666,共4页journal of Computer Applications

基　　金：国家自然科学基金资助项目(30300088)

摘　　要：针对视频序列图像中的运动目标分割,提出了将马尔可夫随机场模型和活动轮廓模型相结合的运动目标分割算法。该算法首先利用马尔可夫随机场模型的运动检测算法,得到运动目标的初始模板。在此基础上提取出活动轮廓模型的初始轮廓点,然后构造活动轮廓模型的能量函数。用改进的贪婪算法求得能量函数最小值,提取出运动目标的精确轮廓,从而得到具有精确边缘的运动目标。实验结果表明该算法能有效地分割和提取出视频序列中的运动目标。A method extracting moving object was proposed by utilizing the Markov random field （MRF） and active contour model in video sequences. Firstly, a rough motion template was obtained by motion detection based on Markov random field model and through post-processing. Secondly, after extracting the initial contour of the object, the energy function of active contour model was constructed. Finally, the solution of minimizing the energy function was obtained by using the modified greedy algorithm, after which, the precise contour of the object was extracted. The experimental results show that the proposed algorithm can effectively segment and extract moving object in video sequences.

关 键 词：活动轮廓模型 马尔可夫随机场 视频分割 目标提取 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]