在小波域进行图像的最大熵分割的一种方法

A Method Applying Maximum Entropy Image Segmentation in Wavelet Field

最大熵分割是一种数字图像阈值分割方法,它以有关的信息理论为基础.本文提出一种新的应用于小波域的最大熵分割方法.这种方法借助"零树编码"的思想,首先在小波变换的低频域进行最大熵分割操作,然后依据这个最大熵分割的结果,我们就能获得原始图像的最终分割结果.实验结果证明该方法与传统的图像最大熵分割方法相比,有远算速度快、抗噪能力强的优点.

Maximum entropy segmentation is a thresholding segmentation method used in digital image processing.It is based on relatively information theory.This paper presents a new maximum entropy segmentation method applied in wavelet domain.With the help of 'zero-tree' thinking,this method does maximum entropy segmentation in the low frequency domain of wavelet transform.Based on the result of the maximum entropy segmentation we can get the final segmentation result of the orignal image.This method is proved by experimental results to have higher computering speed and better denoising performance than traditional segmentation.