Improved SURF algorithm based on SVM classification

An new SURF algorithm based on support vector machine is presented in order to solve the problem of mismatch between feature points. Put the data of the normalized Euclidean distance of feature points into support vector machine to achieve adaptive match after trainning SVM by data. The experiment by OpenCV library verify that the improved SURF algorithm proposed in this paper has higher accuracy than the old one. Besides, there is no significant increase in complexity.