基于最大互信息量的图像自动优化分割

由于传统的阈值分割算法只考虑到图像的灰度信息,而忽略了灰度的空间分布以及分割后图像与原图像之间的关系,因而分割效果不好。为了提高分割效果,从分割图像与原图像的内在联系出发,提出了一种新的基于K均值算法与互信息量(mutual information,MI)技术相结合的分割算法。新算法首先利用K均值算法确定全局阈值作为初值;然后以互信息量为目标函数,在小范围内计算分割图像与原图像的互信息量,互信息量达到最大时的阈值即为最优值。这是将图像配准方法用于分割的一种创新性尝试。通过对大量医学图像以及汽车牌照图像进行的实验结果表明,该新算法所得到的目标图像的边界特征保持完好,不仅虚假目标信息大大降低,而且图像边界细腻、连续,且定位性能好。

Unsupervised Segmentation of Medical Image Based on Maximizing Mutual Information

Most threshold-based segmentation algorithms rely on the information of the gray level of the original image,without taking account of the spatial information. In this paper a new segmentation method is proposed,in which K-means algorithm is combined with mutual information (MI) technique. The initial threshold can be chosen by using K-means algorithm,and in the iteration process,an optimal threshold will be determined by maximizing the MI between the original and the segmented image. We evaluate the effectiveness of the proposed approach by applying it to the segmentation of medical images and license plate images. The experimental results indicate that the new method has visually better segmentation effect.