【摘要】 针对现有肝功能分级方法存在有创性、时效性等问题,提出一种基于CT图像多尺度方向数值模式的肝功能分级方法。利用Gabor滤波器提取肝脏感兴趣区域的多尺度纹理特征,对各尺度的主要方向进行数值模式的紧凑编码并分块统计直方图,获取多尺度特征向量,利用支持向量机构建肝功能分级模型。临床数据实验结果表明,该方法有效且可行,具有非侵入性、高效性和可重复性,为患者肝功能评估提供了基于影像学的辅助诊断。更多还原

【Abstract】 Aiming at the problems of the existing methods of liver function classification,such as the originality and timeliness,a novel method based on multi-scale direction number of CT image was presented,Gabor filters were used to extract the multiscale texture features of the region of interest of the liver.The main dircctions of each scale were encoded in a compact numerical modo and multi-scale feature vectors were obtained by block statistical histogram of coded image in each scale.The liver function was cstablished by support vector machine(SVM).Results of experiments with clinical data show that the proposed method is effective,feasible,non-invasive,efficient and repeatable,which provides imaging-based auxiliary diagnosis for the liver function assessment.更多还原