基于磁共振图像纹理特征提取和分类算法的肝硬化诊断方法。

Diagnostic Method of Liver Cirrhosis Based on MR Image Texture Feature Extraction and Classification Algorithm.

机构信息

Department of Radiology, Jinan Infectious Disease Hospital affiliated to Shandong University, Jinan, 250021, Shandong, China.

Department of Pharmacy, Jinan Infectious Disease Hospital affiliated to Shandong University, Jinan, 250021, Shandong, China.

出版信息

J Med Syst. 2019 Dec 5;44(1):11. doi: 10.1007/s10916-019-1508-x.

DOI:10.1007/s10916-019-1508-x

PMID:31802238

Abstract

In order to improve the accuracy of cirrhosis staging diagnosis based on MR images, a diagnostic method combining image texture feature extraction and classification algorithm is proposed. Firstly, the liver MR image is preprocessed, the region of interest (ROI) image patch is extracted therefrom, and the ROI image is quantized and compressed by the Lloyd algorithm. Then, the ROI image is filtered by a local binary pattern (LBP) operator, and then the texture feature of a 20-dimensional gray-level co-occurrence Matrix (GLCM) in four directions on the LBP image is extracted. Finally, MR image is classified by performing support vector machine (SVM) and the final diagnosis of liver cirrhosis is obtained. The experimental results show that the proposed method can accurately diagnose liver cirrhosis.

摘要

为了提高基于磁共振图像的肝硬化分期诊断的准确性，提出了一种结合图像纹理特征提取和分类算法的诊断方法。首先，对肝脏磁共振图像进行预处理，从中提取感兴趣区域（ROI）图像块，并使用 Lloyd 算法对 ROI 图像进行量化和压缩。然后，使用局部二值模式（LBP）算子对 ROI 图像进行滤波，然后提取 LBP 图像上四个方向的 20 维灰度共生矩阵（GLCM）的纹理特征。最后，通过支持向量机（SVM）对磁共振图像进行分类，从而得到肝硬化的最终诊断。实验结果表明，所提出的方法可以准确地诊断肝硬化。