摘要
应用小波变换对腹部MRI进行图像压缩与平滑化处理,应用边缘检测技术进行边缘数据与肝脏特征值的提取,并结合医师确定的肝脏区域进行图像分割效果的评价。最后利用判别分析法根据特征值进行肝硬化的自动识别。通过对日本国歧阜大学附属医院随机抽取的30枚腹部MRI进行处理,以10%的误判率结果说明了方法的有效性。
?Wavelet transforms are used to compress image and smooth edge of liver region from abdomen MRI. The results of image segmentation are evaluated based on the liver region determined by doctor. Finally, discriminatory analysis method is applied to recognize the image of liver cirrhosis according to the characteristic values. The new technique has been made image recognition of 30 abdomen MRI images that are randomly abstracted from the affiliated hospitals of Gifu University in Japan and got the results of 10?% error rate, which also proved the effectiveness of this method.
出处
《控制工程》
CSCD
2003年第4期360-363,共4页
Control Engineering of China
基金
国家"九五"重点科技攻关项目(97-562-02-02)
