Summary

In this paper an implementation of a feature-preserving volume filtering method has been presented. This method is based on a minimization of a three-component penalty function, so that approximation of the original values, feature preservation and curvature minimization can be controlled efficiently. The scalability is ensured by the weighting parameters of the three-component penalty function. Due to the applied FFT method filtering is performed efficiently. We have implemented a tool, written in Java, to analyze this method by setting various weighting parameters and comparing it to other smoothing methods. Also a C version has been developed to achieve the best computational time.