A new approach to diagnosing prostate cancer through magnetic resonance imaging.

This paper combines improved GrowCut and Zernik feature extraction and ensemble learning techniques such as KNN, SVM, and MLP algorithms to prostate cancer detection and lesion segmentation in MRI. We use improved GrowCut algorithm for segmentation of the suspected cancer area and the combination of machine learning algorithms such as KNN, SVM, and MLP in the ensemble learning system to detect prostate cancer. We found that the accuracy of this method, which is a combination of several methods, improved by about 20% compared to other methods. Other metrics such as precision, recall, and error of proposed method have been improved. [ABSTRACT FROM AUTHOR]