Research on Liver B Ultrasonic Images Classification Based on SVM
Danhong Zhang, Qingyong Zhang · 2010
To improve the correct rate of liver B ultrasonic image classification, a method based on support vector machines is proposed. The gray level co-occurrence matrix is calculated to get the texture feature of liver B ultrasonic image. Then classification using the proposed method with different kernel functions is carried out. The classification results show that RBF kernel can give better performance in most classification groups. Further, a classification system of liver B ultrasonic image is developed by using Microsoft Visual C++6.0. It has good practical application value.