论文部分内容阅读
采用分形模型对超声图像进行了纹理分析，提出了肝脏超声图像具有分形的特点，结合肝脏两类不同的散射体解释了超声图像的二重分形特性，并对十例正常人、三例肝癌患者、三例肝硬化患者肝脏的临床超声图像计算了分维Ｄ，结果表明分维Ｄ是区分病变的重要特征量。根据分维Ｄ 和尺度ε的综合影响，进一步提出纹理分类参数Ｓ作为区分病变的定量指标，实验结果表明参数Ｓ可以有效地区分病变。
The fractal model was used to analyze the texture of the ultrasound images. The fractal features of the liver ultrasound images were proposed. The two fractal characteristics of the liver were interpreted according to the fractal characteristics of the liver. Ten normal and three HCC patients , Three cases of liver cirrhosis patients with clinical ultrasound image calculated fractal D, the results show that the fractal dimension D is an important feature to distinguish lesions. According to the combined effect of fractal dimension D and scale ε, the texture classification parameter S is further proposed as a quantitative index for distinguishing the lesions. The experimental results show that the parameter S can effectively distinguish the lesions.