A Comparative Framework of Probabilistic Atlas Segmentation Method for Human Organ's MRI

Recently, different image analysis methods are used for human body parts. But the internal pectoral muscle segmentation of important body parts in a automatic way is widely used. This is also vital for multi modal image registration. Previously, breast MRI image analysis by automatic pectoral muscle segmentation is studied. In this paper, we introduce a comparative framework of probabilistic atlas segmentation method for breast with brain, chest, heart and liver MRI. For breast, brain, heart and liver and chest segmentation, the obtained DSC values are 0.76 +/- 0.12, 0.71 +/- 0.15, 0.66 +/- 0.08, 0.77 +/- 0.12 and 0.72 +/- 0.13 respectively. The total overlap values for each case are 0.76 +/- 0.12, 0.76 +/- 0.15, 0.71 +/- 0.08, 0.70 +/- 0.12 and 0.70 +/- 0.13 respectively.