A performance evaluation method for geometry-driven diffusion filters

A novel quantitative method is proposed for the algorithm evaluation for geometry-driven diffusion (GDD) filtering methods. It is based on a probabilistic model of stepwise constant image corrupted by uncorrelated Gaussion noise. The maximum likelihood estimates of the distribution parameters of the random variable derived from intensity gradient are used for characterization of staircase image artifacts in diffused images. The proposed evaluation technique incorporates a "gold standard" of the GDD algorithms, defined as a diffusion process governed by ideal values of conductance. A phantom mimicing an MR brain scan is used as a sample data set.