分解的二维非对称Tsallis交叉熵图像阈值选取

Image Threshold Selection Based on Two-Dimensional Asymmetric Tsallis Cross Entropy and Decomposition

现有的Tsallis交叉熵能够度量图像分割前后的差异,但公式复杂,计算效率不高,据此,提出了基于分解的二维非对称Tsallis交叉熵图像阈值选取方法。首先给出了非对称Tsallis交叉熵的定义,提出了一维非对称Tsallis交叉熵阈值选取方法;然后,将其拓展到二维,推导出相应的阈值选取公式;最后,在此基础上提出了二维非对称Tsallis交叉熵阈值选取的分解算法,使求解二维非对称Tsallis交叉熵阈值法的运算转化到两个一维空间上,将计算复杂度从O(L4)降低为O(L)。大量实验结果表明,与基于混沌粒子群优化的二维Tsallis灰度熵法、二维斜分对称交叉熵法,二维斜分对称Tsallis交叉熵法等方法相比,该方法分割性能优,运行时间短,可望满足实际应用系统对分割的实时要求。

The existing Tsallis cross entropy can measure the difference between the original image and its segmentation result, but it has the drawback of complex formula and low computational efficiency. Thus two-dimensional asymmetric Tsallis cross entropy threshold selection method based on decomposition is proposed. Firstly, the asymmetric Tsallis cross entropy is defined and a one-dimensional threshold selection method based on the asymmetric Tsallis cross entropy is put forward. Then it is extended to the two-dimensional space, and the corresponding threshold selection formulae are derived. Finally, the decomposition algorithm of two-dimensional asymmetric Tsallis cross entropy thresholding is proposed on this basis. As a result, the computations of two-dimensional asymmetric Tsallis cross entropy thresholding method are converted into two one-dimensional spaces. The computational complexity is greatly reduced from O（L4） to O（L）. A large number of experimental results show that, compared with two-dimensional maximum Tsallis gray entropy method based on chaos particle swarm optimization, symmetric cross entropy method based on two-dimensional histogram oblique segmentation, symmetric Tsallis cross entropy method based on two-dimensional histogram oblique segmentation and so on, the proposed method has superior image segmentation performance and short running time, which can meet the real-time processing requirement of segmentation in the practical application systems.