Brain Segmentation using ATP (Automatic Twice PAM) in Multi Diffusion Indices
Chao Y, Cho K, Chen D, Chen J, Yeh C, Lin C, Tsao S
Institute of Electrical Engineering, National Taiwan University
Non-invasive diffusion tensor imaging (DTI) has been used widely in brain research and clinical applications recently. Various diffusion indices were developed for recognizing the constituents of brain and the nidus of diseases. In this research, Automatic Twice PAM (ATP) clustering algorithm was employed to segment the brain architectures with features of common diffusion indices with no initial information. The segmentation was found to agree strongly with the anatomy atlas. This efficient and robust method also showed the outstanding ability for aiding morphological study and analysis of brain development.