蜂群优化的二维非对称Tsallis交叉熵图像阈值选取

Two-dimensional asymmetric tsallis cross entropy image threshold selection using bee colony optimization

交叉熵能够度量图像分割前后的差异,与Shannon交叉熵相比,引入参数q的Tsallis交叉熵则为图像阈值分割提供了灵活性和普适性,而非对称Tsallis交叉熵的表达形式更加简洁。由此,提出了蜂群优化的二维非对称Tsallis交叉熵图像阈值选取方法。首先引出了非对称Tsallis交叉熵,导出了二维非对称Tsallis交叉熵阈值选取公式,并利用递推方式计算阈值选取准则函数涉及的中间变量,建立查找表,消除冗余运算;然后采用蜂群算法搜寻最佳二维阈值。大量实验结果表明,相对二维最大Shannon熵法、二维Shannon交叉熵法、二维Tsallis熵法和二维对称Tsallis交叉熵法等同类方法,所提出方法在主观视觉效果和区域间对比度评价指标上有较大的改善,能够更准确地分割出目标,运行速度也更快。

Cross entropy can measure the difference between the original image and its segmentation result. Compared with Shannon cross entropy,Tsallis cross entropy,in which a parameter q is introduced,provides flexibility and universality for the segmentation of image threshold. The asymmetric Tsallis cross entropy has more concise expression form. Therefore,a method of threshold selection is proposed based on the two-dimensional asymmetric Tsallis cross entropy using bee colony optimization. Firstly,the asymmetric Tsallis cross entropy is introduced and the threshold selection formulae based on the two-dimensional asymmetric Tsallis cross entropy are derived. Recursive algorithms are used to calculate the intermediate variables involved in criterion function for threshold selection and a lookup table is built to eliminate the redundant operations. The optimal two-dimensional threshold is searched by the bee colony algorithm. A large number of experiment results showed that the proposed method is greatly improved in terms of subjective visual effect and inter-regional contrast evaluation indicators compared to the relevant methods,such as the two-dimensional maximum Shannon entropy method,the two-dimensional Shannon cross entropy method,the two-dimensional Tsallis entropy method,and the two-dimensional symmetrical Tsallis cross entropy method.It can segment objects more accurately and has a faster running speed.