Research on Segmentation Method of Multi-Region Liver Image
Huiyan Jiang, Hiroshi Fujita · 2006
A new segmentation method of multi-region liver image is suggested in this paper. In the new method, the aorta was segmented based on improved Ostu algorithm and its shape characteristic, then the liver image in single regions was segmented based on the position of aorta, region growth method of multi-seed voting mechanism and self-adaptive adjusting threshold. Finally, multi-region liver images were carried out by extracting candidate region of liver based on its grey scale and position, and matching and merging the segmented liver regions. The new method was applied to the actual 30 pieces MRI and show its availability to multi-region segmentation of liver images.