Clayton Chi-Chang Chen1, Jyh-Wen Chai1, San-Kan Lee1, Yen-Chieh Ouyang2, Chein-I Chang3, Wu-Chung Shen4, Hsian-Min Chen4
1Department of Radiology, Taichung Veterans General Hospital, Taichung, Taiwan; 2Department of Electrical Engineering, National Chung Hsing University, Taichung, Taiwan; 3Remote Sensing Signal and Image Processing Laboratory, Department of Computer Science and E.E., University of Maryland, Baltimore County, USA; 4Department of Radiology, China Medical University Hospital, Taichung, Taiwan
Independent component analysis implemented with support vector machine has the advantages of using an unsupervised technique to separate the distinct objects and then followed by a supervised classification technique to perform target substance discrimination. The method could be effective in image analysis of the major components of normal and diseased brain in multispectral MRI. However, there was a lack of comprehensive assessment of the proposed method for brain segmentation in the clinical applications. In this study, we tried to carry out an experiment to test the accuracy and reproducibility of the proposed method in the synthetic and clinical MRI data.