CAJ | 학술논문

由于医学X光图像中存在信噪比低、清晰度差、对比度低等缺点，而普通的图像增强算法很难在增强图像细节特征的同时对图像的背景噪声进行抑制。针对上述问题，将模糊多尺度Retinex算法引入医学图像处理的增强算法中：先计算图像中每个像素点的模糊度，再结合模糊度采用多尺度Retinex算法对图像进行增强。实验表明该算法比直方图均衡、同态滤波以及普通的多尺度Retinex算法增强图像的效果更明显，并能在增强图像的同时，有效抑制图像的背景噪声。
유우의학X광도상중존재신조비저、청석도차、대비도저등결점，이보통적도상증강산법흔난재증강도상세절특정적동시대도상적배경조성진행억제。침대상술문제，장모호다척도Retinex산법인입의학도상처리적증강산법중：선계산도상중매개상소점적모호도，재결합모호도채용다척도Retinex산법대도상진행증강。실험표명해산법비직방도균형、동태려파이급보통적다척도Retinex산법증강도상적효과경명현，병능재증강도상적동시，유효억제도상적배경조성。
The medical X-ray radiography images always have the character of low Signal-Noise Ratio ( SNR) ,bad definition and con-trast,and it is difficult to reduce the noise while enhancing the detail characters by the custom image enhancement algorithm. According to the problems mentioned above,Fuzzy Multi-Scale Retinex ( FMSR) algorithm was applied to medical image enhancement. Firstly the fuzzy degree of each pixel was calculated,then the MSR algorithm was applied combined with fuzzy degree to enhance the image. Experi-ment shows that compared with the custom image enhancement algorithm as histogram equalization,homomorphic filtering and the cus-tom MSR algorithm,this algorithm can get a better performance,and the noise can be efficiently reduced while enhancing the image.