Application of Multi-Classification Support Vector Machine in the B-Placenta Image Classification
Zhi Liu, Herong Zheng, Shengliang Lin · 2009
In this paper, B-placenta image is classified automatically using support vector machine based on feature extraction. Firstly, artificial selected region of interest (ROI) is regarded as the object of feature extraction. Then traditional gray-scale statistical analysis is used to extract the characteristic parameters of B-placenta image as the basis data for the placenta classification. The binary tree multi-classification SVM is used to automatically classify the B-placenta image. The binary tree generation algorithm is optimized based on the ultra-radius. The experiment shows that this classifying method using binary tree multi-classification SVM in the B-placenta image classification has very high value.