基于分形参数的B超图像纹理分析

Texture Analyses of B Scan Image Based on Fractral Dimension

利用离散分形布朗随机场（ＤＦＢＲＦ）模型对Ｂ超图像的分形特征进行了分析和讨论．鉴于一般的图像不是完全的分形体，文中给出了计算分维数的尺度范围公式．根据上述模型和尺度范围公式，对正常肝脏和有癌变肝脏的两组Ｂ超图像的分维数进行了计算，结果表明在未考虑尺度范围时两组图像的分维数差别不大；而考虑了尺度范围时两幅图像的分维数就有明显的差别，即正常肝脏图像的分维数大于有癌变肝脏图像的分维数．

The fractal feature of the B scan images is analyzed and discussed using the discrete fractional Brownian random field (DFBRF) model. Since a general image is not a complete fractal object, a scale range formula is used to calculate fractal dimension. According to the DFBRF model, the fractal dimensions of two groups of B scan images are computed. When a scale range is considered, fractal dimensions of two groups of B scan images show an appreciable difference. That is, the fractal dimension of an image for a normal liver is higher than that of a liver cancer. If the scale range is not taken into consideration the difference is small.