中文摘要：

提出一种基于改进型的自适应分数阶微分的B超图像增强方法。该算法根据B超图像的特点,在分数阶微分定义的基础上对分数阶微分掩模算子分解的8个方向的模板进行改进,得到像素点周围8个方向的加权求和值,取其最大值作为增强后的像素值,再利用伪彩色变换算法进一步增强。与常用的灰度变换、均值滤波等B超图像增强方法相比较,结果表明：基于改进型的自适应分数阶微分能很好地突出图像的边缘和纹理,抑制较大噪声,结合伪彩色变换能较准确地从高强度聚焦超声（HIFU）辐照后的B超图像中检测出治疗区域,与切片实验所获取的治疗区域大小、形状更趋一致,能有效地帮助临床医生确定治疗区域的大小、形状及位置。

英文摘要：

An enhancement method of B-mode ultrasonic image based on the improved adaptive fractional-order differential is presented.According to the characteristics of B-mode ultrasound images,the operator of fractional-order differential is divided into eight sub-templates with different directions around the detecting pixel,and then the eight weight sum values for the eight sub-templates are obtained.Then a maximum of eight directions is given to the original pixel.What＇s more,the pseudo color transform algorithm is used to strethen further.The result shows that the improved fractional-order differential can enhances the image edges and the texture details,and avoid loud noise more effectively,compared with commonly used gray level transformation,median filter and ultrasound image enhancement methods.Combined with pseudo color transformation,the algorithm can detect the treatment area precisely from B-mode ultrasound image after irradiating by HIFU,which is consistent with practical size and shape of the treatment areas obtained by slicing up the irradiated fresh pork.It can help clinicians to observe and identify the size,shape and position of the treatment area effectively.