Sigmoidal Weighted Vector Directional Filter

摘要

In this paper, we provide and analyze a sigmoidal optimization of a recently developed class of weighted vector directional filters (WVDFs) output-ting the input multichannel sample associated with the minimum sum of weighted angular distances to other input samples. Because the WVDFs can perform a number of smoothing operations in dependence on the weight coefficients, the aim of this paper is to adapt the WVDF behavior to statistical properties of noise and original color image. The filtering results and the complete analysis of the sigmoidal function based WVDF optimization are also provided.