Hemodynamic segmentation of MR brain perfusion images using Independent Component Analysis, Thresholding, and Bayesian Estimation (ICA-TBE) techniques
Yi‐Hsuan Kao, Weidong Guo, Yiming Wu, K. C. Liu, Weilong Chai, Chuang Lin, Y S Hwang, J. K. Liou, Hsiu‐Mei Wu, Hang Cheng, Ta‐Chuan Yeh, Jen‐Chuen Hsieh, Ming Teng · 2003
Synopsis Dynamic-susceptibility-contrast perfusion imaging is a widely used imaging tool for in vivo study of cerebral blood perfusion. However, visualization of different hemodynamic compartments is less investigated. We use independent component analysis, thresholding, and Bayesian estimation to concurrently segment different tissues on perfusion images of normal volunteers. Based on the spatio-temporal hemodynamics, sequential passages and microcirculation of contrast-agent particles in brain tissues are decomposed and analyzed. Late and multi-phasic perfusion, indicating the presence of contrast agents, is observed in choroid plexus and CSF areas. An arterial input function is modeled from an arterial area on the same slice.