Multi-features prostate tumor aided diagnoses based on ensemble-svm

In order to realize prostate cancer aided diagnosis, an ensemble SVM which based on kernel functions and feature selection is proposed. Firstly statistical, texture and invariant moment features of the prostate ROI in the MRI images are extracted. Secondly SVM parameters are disturbed by different kernel functions in different features space, and the first integration is carried out by relative majority voting. Thirdly the first results of ensemble are integrated by relative majority voting again; Finally, MRI images of prostate patients are regarded as original data, and the new ensemble SVM is utilized to aided diagnosis. Experimental results show that the proposed algorithm can effectively improve the recognition accuracy of prostate cancer.