Automatic Brain Segmentation from Head MRI Data using Region Growing Method with Restricted Processing Area

I. Uwano, K. Masashi, K. Matsuda, A. Doi, T. Inoue, and A. Ogawa (Japan)


Medical Image Processing, MR images, Segmentation, Region Growing,


The Region Growing method, which is used in most of the segmentation, has problems that a leakage and a few isolated regions are generated in 2-dimensional domain. Therefore, we propose an efficient brain segmentation method for improving those problems. Our method has four step : 1) edge detection, 2) selecting ROI, 3) Region Growing with restricted processing area every slice, 4) two-way Region Growing. We applied our method to the T1-weighted MR image data of a patient. As the result, we have improved the leakage and the insufficiency of the extraction of brain data.

Important Links:

Go Back