Blind Source Separation of Hemodynamics from Magnetic Resonance Perfusion Brain Images Using Independent Factor Analysis
Figure 3
Classification results of a normal subject (whose perfusion images are shown in Figure 1). (a) Five major tissue types exhibited in IF images are ar, GM, WM, vs, and cp, respectively. (b) The normalized signal-time curves correspond to five IF images, where the one (red) with the fastest signal drop is the artery. (c) Left: a color-coded composite map is used to represent the final segmentation result for different tissue types. Right: five signal-time curves correspond to the averaged intensities of color-coded areas. (d) The rCBV map (left), the rCBF map (middle), and the MTT map (right). (Scale unit for rCBV and rCBF: arbitrary unit; scale unit for MTT: second.)
