快速二维最小交叉Tsallis熵的图像阈值分割

摘要

The current thresholding method based on 2-D minimum cross Tsallis entropy has good segmentation performance, but owing to high complexity its speed is so slow. So a fast image thresholding method based on 2-D minimum cross Tsallis entropy is presented. Firstly, the formula of 2-D minimum cross Tsallis entropy is deduced to find some variables to be recurred. Then, a two-dimensional histogram is analyzed to get its features.Finally, a new recursive approach is inferred with the features to reduce the computational complexity. Experimental results show that the proposed method' s computing time is less than O. 2 second and that its running speed is over 20 times faster, with the same segmentation result, than that of the current thresholding method based on 2-D minimum cross Tsallis entropy.%目前二维最小交叉Tsallis熵阈值分割法有较好的分割性能,但由于计算复杂度高,使得分割速度慢.针对此问题,提出了一种基于二维最小交叉Tsallis熵的快速图像分割方法.首先对二维最小交叉Tsallis熵法公式进行推导找出需要递推的几个量,然后对二维直方图投影进行分析得到二维直方图的特性;最后利用此特性导出新型的快速递推算法来减少计算时间.实验结果表明:相对于当前二维最小交叉Tsallis熵阈值法,提出的方法在保持分割效果的情况下,其速度提高了20倍以上,其运行时间小于0.2 s.