基于改进混合遗传的医学图像模糊增强研究*

针对现有医学图像中存在有采集后图像质量不高、图像过暗等现象,对遗传算法中的选择、交叉、变异特性进行研究,同时结合粒子群优化、禁忌搜索及模糊增强算法,提出一种基于改进混合遗传的医学图像模糊增强方法.该方法通过对传统遗传算法改进,将粒子群优化思想及粒子空间对称分布原理引入以改善遗传算法缺乏明确的目标指向性、“突变”性过高的现象,并且为有效降低粒子的同一位置二次搜索,在算法执行过程中加入了禁忌搜索算法.最后,通过与模糊增强算法相结合,并设置二维方向寻优,可自适应的同时寻找到两个模糊参数Fp、Fe最优值,完成医学图像的模糊增强.实验结果表明,改进后算法可有效改善过暗医学CT图像的质量,增强效果较好.

Research on an medical image fuzzy enhancement based on improved hybrid genetic algorithm*

Aiming at the phenomenon of existing medical images that the image quality is not high or too dark when collected,re searching for the selection,crossover and mutation characteristics of genetic algorithm and combined with particle swarm optimization,tabu search,fuzzy enhancement algorithm that an medical image fuzzy enhancement method based on improved hybrid genetic algorithm is proposed.In this method,improved traditional genetic algorithm by introducing the thought of particle swarm optimization and particle spatial distribution principle,so the situation is improved that of genetic algorithm was lack of clear-oriented and 'changing'was too high.And adding tabu search algorithm in order to reduce the duplication optimization of particles by recording the position of the particle search in the later of iteration.Finally,the algorithmic search direction are setted to the 2D and combined with fuzzy enhancement algorithm that can adaptive search the optimal value of fuzzy parameters Fp and Fe to achieve medical image fuzzy enhancement.The experiments show that the improved algorithm is better for the enhancement of medical CT image that was too dark and can effectively improve the image quality.