基于图像熵的快速Chan-Vese模型分割算法

提出了基于图像熵的快速Chan-Vese模型分割算法.该算法利用实时图像熵自适应计算模型能量函数中的拟合参数以提高分割速度,并通过检测熵在曲线形变过程中的变化来判定曲线演化的稳定态.实验表明.针对含噪严重、目标模糊且边缘不连续的红外图像目标检测,所提出的分割算法可以取得精确、高效的分割结果.

Fast ChanVese Segmentation Algorithm Based on Image Entropy

The Chan-Vese model based on level set method is more robust than other curve evolution models in detecting objects with discontinuous edges under noisy environment. Even though it has been widely used in the field of image segmentation, it has several problems such as parameter chosen, segmentation speed and iteration termination. In order to alleviate these problems, a fast Chan-Vese segmentation algorithm based on image entropy is presented. The image entropy was computed as the fitting term parameters to speed up the evolution. Meanwhile, the evolutionary stability was determined by analyzing the change of the entropy. The experiments on a variety of infrared images show its superiority in effectiveness and efficiency on noisy, object blurred and edge discontinuous image detection.