基于分形小波变换的医学图像增强  

Enchantment in medicine image based on fractal-wavelet transform

作　　者：李清顺[1] 杨定楚[2] 秦前清[1] 

机构地区：[1]武汉大学测绘遥感信息工程国家重点实验室,湖北武汉430079 [2]武汉大学电子信息学院,湖北武汉430072

出　　处：《计算机工程与设计》2005年第3期807-809,共3页Computer Engineering and Design

摘　　要：Mandelbrot 的分形理论认为,自然界中的大量图像都可以用分形布朗运动模型来描述,而分形维数可以表征图像自相似性和灰度表面粗糙度。充分地利用图像的分形维数的特性和小波变换的边缘保持好、适合人眼的特点来增强图像,辅助医学临床诊断。通过对X 光胸片图像的增强实验表明:该方法不仅能增强图像中较不敏感的细节内容,还能增强视觉上很难感受到的差异结构。According to mandelbrot's theory, the nature images can be regarded as the result of FBM (fractional brownian motion) and fractal dimension can describe excellently the character of image self-similitude and coarseness of its gray surface. A method of image enchantment is proposed based on fractal & wavelet theory which can be used to assist clinic medical diagnosis. Experiments on x-ray sternum images show that the method can strengthen the non- sensitive parts in image and differential structures which is hard to distin- guish for human vision.

关 键 词：诊断 X光胸片 医学图像 辅助 医学临床 人眼 实验 边缘保持 小波变换 分形布朗运动 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]