Renyi熵与Tsallis熵的等价关系

A New Image Threshold Segmentation Algorithm

描述了Tsallis熵和Renyi熵的等价关系及其在图像分割领域的应用。Tsallis熵和Renyi熵在图像分割中的应用是两种重要的全局阈值选取方法,是Shannon熵的广义形式,不仅在形式上具有等价关系,而且在应用于图像分割中也具有特殊的等价关系。这种特殊的等价关系应用在图像分割时只需任选一种熵来选取阈值。还讨论了它们自带的参数的取值对阈值选取的影响,这个参数可以针对不同类型的图像进行调节,可以得到很好的分割效果。

The equal relationship of Tsallis＇ s and Renyi＇ s entropy is discussed and these entropies are applied in image segmentation. Tsallis entropy based and Renyi based image segmentation are two important global threshold selection approaches in image segmentation. They the generalization of Shannon entropy. They show the equal relationship not only in formalism but also in the application in image segmentation. The relationship is proved. This special relationship can select any of them to gain the threshold value when they are applied in segmentation. The parameter of these entropies affects the threshold and can be adopted to different types of images. The segmenting results are good.