AN ALGORITHM TO SEGMENT THE MID-BRAIN STRUCTURES USING MULTI- RESOLUTION NON-RIGID REGISTRATION


 In present study, a reference based technique has been proposed to segment mid- brain structures in 3D Magnetic Resonance (MR) images using multi-resolution non-rigid registration. The scheme segmented the mid-brain two stages. During first stage, target image was registered an already while second segmentation information mapped from medical image. Using this scheme, aim nine including left ventricle, right third anterior commissure, posterior putamen, caudate nucleus and nucleus. Results of automatic achieved through registration were validated by comparing with manual segmentation, considered as ground truth. Three metrics namely Dice Coefficient, Sensitivity Positive Predictive Values used for assessing results. algorithm showed satisfactory achieving mean sensitivity 89. 85 % five MR images. results indicated that performed accurate structures.