马尔可夫边缘描述符的图像特征提取算法【荐】.doc

马尔可夫边缘描述符的图像特征提取算法 摘要:分析目前mpeg.7标准下对纹理特征提取的研究，在边缘直方图描述符的基础上，借鉴马尔可夫链的思想，提出了马尔可夫边缘描述符。该描述符使用马尔可夫链的初始分布表征同种类型边缘之间的空间分布信息，采用马尔可夫链的极限分布来表征不同类型边缘之间的空间分布信息。实验表明，马尔可夫边缘描述符在具有直方图描述符优点的基础上，可以很好地捕获图像边缘的空间信息，对同种或不同种边缘类型间的空间分布都有较好的检索效果，检索效果优于边缘直方图描述符。 关键词:基于内容的图像检索；特征提取；mpeg.7；马尔可夫链；边缘描述符 markov edge descriptor for feature extraction algorithms of image  cui ning.hai*, liu li.ping college of information science and engineering, shenyang ligong university, shenyang liaoning 110159, china abstract: after analyzing the texture feature extraction on mpeg.7 standard, a new markov edge descriptor (med) was proposed base on edge histogram descriptor (ehd), which was in virtue of the markov chain. it adopted initial distribution of markov chain model to represent the spatial information between the same kind of edges. and it adopted stationary distribution of markov chain model to represent the spatial information between the different of edges. experimental results show that medhas the merits of ehd, and med can describe the spatial information between the image edges. it has corking retrieval performance for the same or different kind of edges. its retrieval performance is better than ehd. after analyzing the texture feature extraction on mpeg.7 standard, a new markov edge descriptor (med) was proposed based on edge histogram descriptor (ehd), which was in virtue of the markov chain. it adopted initial distribution of markov chain model to represent the spatial information between the same kinds of edges. and it adopted stationary distribution of markov chain model to represent the spatial information between different edges. the experimental results show that med has the merits of ehd, and med can describe the spatial information between the image edges. it has corking retrieval performance for the same or different kinds of edges. its retrieval performance is better than ehd. key words: content.based image retrieval; feature extraction; mpeg.7; markov chain; edge descrip