MRI中的肿瘤边缘的自动检测  

Automatic Edge-Detection of Tumor in MRI

作　　者：袁华[1] 吴效明[1] 岑人经[1] 袁支润[2] 

机构地区：[1]华南理工大学生物力学研究所,广东广州510640 [2]四川大学生物医学工程研究所,四川成都610065

出　　处：《华南理工大学学报（自然科学版）》2001年第4期6-8,共3页Journal of South China University of Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目! (19972 0 44 );广东省自然科学基金资助项目! (990 6 6 8)

摘　　要：MRI用于临床诊治时 ,肿瘤大小的自动检测一直是尚未圆满解决的问题 ,其中的关键技术是肿瘤边缘的自动检测 .边缘检测是计算机初级视觉中的一个难题 ,一直受到图像处理研究人员的关注 .虽然边缘检测的算子很多 ,但都是在噪声滤除与边缘检出之间作权衡取舍 .本研究中提出了一种结合形态滤波器和广义模糊算子自动检测MRI中的肿瘤边缘的新方法 ,它能在保留边缘细节的同时最大程度地抑制噪声 .实验结果表明 ,在噪声抑制和边缘点检出的折衷中 。Measuring the tumors in MRI automatically has always been a problem and no satisfactory solution has yet been offered. The key technology is the automatic detection of the tumor edges. Edge_detection is also a big problem in computational low_level vision and has always attracted the attention of image_processing researchers. There are many edge_detectors, but they are limited in making a balance between noise_suppressing and edge_detection. This paper presents a new method, which detects the edges of tumor in MRI using morphological filter and generalized fuzzy operator. It can preserve edge details while suppressing the noises. Experimental result shows that the new method does better than traditional methods.

关 键 词：核磁共振成像 形态学滤波器 广义模糊算子 边缘检测 医学图像分析 肿瘤检测 

分 类 号：R445.2[医药卫生—影像医学与核医学] TP391.41[医药卫生—诊断学]