基于模糊神经网络的乳腺微钙化点检测  

Microcalcification Detection Based On Neural Network

作　　者：李清美[1] 

机构地区：[1]泰安市中心医院影像中心,山东泰安271000

出　　处：《中国伤残医学》2009年第1期11-14,共4页Chinese Journal of Trauma and Disability Medicine

摘　　要：在乳腺钼靶x线片计算机辅助诊断中,微钙化点的自动检测是难点问题之一.本文基于模糊神经网络,建议了一种乳腺微钙化点提取的新方法.算法首先利用随机方法产生大量的样本,经模糊判别后,输入神经网络进行训练,训练后的神经网络对感兴趣区域进行分类得到微钙化点.最后,仿真结果同其它算法进行比较,证明本文算法具有更高的阳性检出率。Objective： In the computer--aided diagnosis system of mammogram, automatic detection of microcalcification is one of the most difficult problems. In this paper, we proposed a new method of microcalcification detection based on fuzzy neural network A lot of random training samples were firstly produced; then, these samples were classified into the background and microcalcifiations by using fuzzy method. After training, the neural network was used to classify the regions of interest. Finally experimental results were compared with the results of other algorithm, and demonstrate better true positive detecting ratio.

关 键 词：神经网络 微钙化点 模糊数学 感兴趣区域 

分 类 号：R655.8[医药卫生—外科学]