基于模糊最大熵的医学X增强方射线图像边缘法研究  

Research on Edge Enhancement Algorithm of Medical X-ray Image Based on Maximal Fuzzy Entropy

作　　者：田星[1] 桂志国[1] 刘芳[1] 

机构地区：[1]中北大学电子测试技术国家重点实验室,太原030051

出　　处：《中国医学物理学杂志》2014年第3期4889-4892,共4页Chinese Journal of Medical Physics

基　　金：山西省自然科学基金资助项目(2009011020-2)

摘　　要：目的:由于医学X射线图像在数字化成像过程中容易受到成像设备中射线散射、电器噪声以及人体组织结构的复杂性等因素的影响,导致数字医学X射线图像的质量不高。因此,针对数字医学X射线图像对比度较差,目标细节信息不明显的特点,研究了一种基于模糊最大熵的图像边缘增强算法。方法:首先将医学X射线图像从灰阶域变换到模糊域,然后通过最大熵准则确定模糊阈值将医学X射线图像分为目标和背景两部分,并分别对其进行图像增强处理,最后再映射回到灰阶域。结果:本文以主动脉造影X射线图像为例,对其分别进行经典模糊边缘增强、反锐化边缘增强和模糊最大熵边缘增强处理,并对处理后图像的相关参数进行定量分析。结论:结果表明基于模糊最大熵算法处理后的图像质量高,边缘细节信息明显增强,且该算法相比其它两种算法具有更好的抗噪性。Objective：Due to the X-ray images were susceptible by many factors such as the ray scattering Of relative equipment, electrical noise and the complex structure of human organizational during the digital imaging process, resulting in the low quali- ty of the digital medical X-ray images： In view of the features of d!gital medical X-ray imagel which is poor contrast and unob- vious details, this paper studied a kind of image edge enhancement algorithm based on fuzzy theory. Methods：First taking the image from the gray field to fuzzy field by fuzzy transformation ,second the fuzzy threshold was applied to dividing the image into two parts： target and background, which Were enhanced differently, and the optimal ~eshold was searched through the maximal fuzzy-entropy criterion, last the enhanced image was back to gray filed. Results：In this paper,we take the aortic an- giography X-ray image as example, the classic pal-king algorithm, unsharp mark theory and the fuzzy edge enhancement based on maximum entropy algorithm are respectively apply to the image edge enhancement procegsmg, and comparing related pa- rameters Of the treatment images. Conclusious：Experimental results show that the improved fuZzy image enhancement based on maximum entropy algorithm processed image quality is higher than the other, and the detail information enhanced, and this method has stronger noise resistance.

关 键 词：X射线图像 模糊理论 最大熵 边缘增强 

分 类 号：TP311.52[自动化与计算机技术—计算机软件与理论]