基于GLCM的脑部CT图像纹理特征提取方法研究

Research on Texture Feature Extraction of Brain CT Images Based on GLCM

运用灰度共生矩阵视角下,在Matlab2010平台上研究正常脑部CT图像与异常脑部CT图像的纹理特征值即能量、对比度、逆差分矩、熵之间的差异。结果表明,除对比度特征值外,能量、逆差分矩、熵等3项图像纹理特征值均能有效地区分正、异常脑部CT图像。基于灰度共生矩阵的纹理特征提取方法,用于脑部CT图像模式识别分类特征的提取是可行的。

In the perspective of gray level co-occurrence matrix,the differences of CT image features of the abnormal brain were studied with the help of Matlab2010 platform,which included the differences of energy,contrast,deficit and entropy. The results showed that in addition to the characteristics of the contrast,the energy,the deficit score moment and the entropy can effectively distinguish between positive and abnormal brain CT image. Gray level co-occurrence texture feature extraction method is feasible for brain CT image pattern recognition classification feature extraction.