摘要
提出一种基于区域二维熵的新型主动轮廓模型。该模型将基于二维直方图的熵作为主动轮廓的新型外部能量,能有效完成内部存在噪声和虚假边缘的同质区域的分割。同时针对多数基于区域的主动轮廓模型只分割同质目标的局限,本文通过使用非参数化方法有效表达目标统计特性,并经由模型演变过程中自相交问题的解决,完成了内部结构较复杂的非同质目标(如人脸)的分割。实验验证了该模型的有效性。
On the basis of active contour model , we introduce a new region-based energy criterion- non-parametric 2D entropy. In this new external energy, 2D grayscale histogram is used to obtain the prior probability of the pixels in the image, and non-parametric techniques are used to obtain the conditional probability, which is the element of entropy. This new active contour model can effectively accomplis edges. Furthermore self-intersection, we Experimental results h the segmentation of homogenous objects that contain noise and pseudo , by the non - parametric method and the settlement of the problems with can segment more complex inhomogeneous objects （such as human face）. have shown the effectiveness of the proposed model.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
2005年第6期717-722,共6页
Pattern Recognition and Artificial Intelligence
关键词
主动轮廓模型
二维熵
非参数方法
自相交
图像分割
Active Contour Model, 2D Entropy, Non-Parametric Method, Self-Intersection,Image Segmentation
