Lattice Boltzmann Algorithm for Vessel Enhancement Problems

In this paper, a five-velocity lattice Boltzmann model with multiple non-constant relaxation times is applied for modeling anisotropic diffusion processes with applications to vessel enhancement problems. The proposed algorithm is as follows. First, the Frangi vesselness function is evaluated. In the non-vessel regions a lattice Boltzmann model for the Perona–Malik nonlinear diffusion equation is adopted in order to damp noise. In the regions with vessel-like structures a lattice Boltzmann model for coherence-enhancing anisotropic diffusion process is employed. It is shown in the numerical experiment that the proposed approach is efficient for noise removal, coherence enhancement and eventual binarization of vessel trees in synthetic images.