Development of Ultrasonic Image Processing Systems for Liver Diseases

该研究项目旨在开发肝脏疾病的超声图像处理系统。如果我们使用这些系统，那么就可以自动判断肝脏疾病症状的程度。例如，在所提出的方法中，肝硬化被分为4类。 0型表示肝脏正常，1型表示轻度肝硬化，3型表示重度肝硬化。使用二维 FFT 进行平滑处理、使用数字滤波器进行噪声去除以及阴影补偿均在个人计算机上执行。此外，使用所提出的算法自动确定超声图像中用于特征提取的图像域。各种模拟结果表明，所识别的类别与熟练医生的分类非常吻合。因此，所提出的方法对于实际应用非常有用。

This research project aims at the development of ultrasonic image processing systems for liver diseases. If we use these systems, then the degree of symptoms of liver diseases can be judged automatically. The cirrhosis of liver, for example, is divided into 4 categories in the proposed method. Type 0 means normal liver, and type 1 denotes light degree cirrhosis, and type 3 is heavy one. Smoothing using two-dimensional FFT, noise removal by digital filters and compensation of shading are executed on a personal computer. Furthermore, the image domain for feature extraction in the ultrasonic image is determined automatically by using the proposed algorithm. Various simulation results show that the identified categories agree with those by skillful doctors very well. Thus the proposed method is very useful for practical applications.