Visually controlled on-demand derivation and visualization of perfusion parameters
The visualization of concentration time curves is centered around precomputed descriptive perfusion parameters. Different computational methods of deriving these parameters lead to different results, and there is little consensus of which methods to prefer in different circumstances. In this thesis an interactive visual analysis framework is proposed to derive perfusion parameters on-demand. Using information from a curve view, smoothed concentration time curves and derivatives are generated using Gaussian smoothing. Based on an analysis of the smoothed derivatives using a curve view visualization, perfusion parameters are derived using function fitting, with initial parameters based on the smoothed curves. The proposed framework allows for comparison of different computational methods, as well as insight into the expressiveness of the derived perfusion parameters.