An Unsupervised method for brain MRI Segmentation
Gang Zhang, Dan Zhou, Ying Jeh Huang, Xiaobo Huang, Yong Zhang, Weishi Luo, Yang Zhang, Qian Dongxiang · 2013
Computer aid diagnosis for brain MRI image is widely used in hospitals. An important step in it is to recognize different regions within an MRI image according to medical experience. In this paper, we propose an unsupervised learning algorithm for automatic segmentation of MRI images. Different from previous methods, our method achieves the idea of visual segmentation, which simulates the thinking procedure of doctors. And prior knowledge can be incorporated in our model. Evaluation results on a synthetic brain MRI dataset and a real dataset show the effectiveness of the proposed method.