Blind source separation can be used to linearly decompose DCE-MRI time-course data into a sparse set of time courses, or sources, and maps of coefficients, or weights, to describe the entire 4D dataset. This type of analysis generates in realistic time-courses for the wash-in and wash-out of the contrast agent, and maps of the distribution of these dynamics. In turn, these decompositions may hold diagnostic value. Random initialization typical of such algorithms makes the output unstable. This work sought design an approach to blind source separation analysis of DCE-MRI with lower variability and independent of NMF initialization.
This abstract and the presentation materials are available to members only; a login is required.