This packaged is an implementation of our paper "Robust Denoising of Piece-Wise Smooth Manifolds", ICASSP 2018 The algorithm creates an affinity graph and perform denoising on a set of N input points in R^n. Given an input set of points in any arbitrary dimension, an affinity graph is first created based on Tensor Voting, Local PCA or Euclidean distances, or the Tensor Voting Graph [3] . Then it performs denoising using a modified version of the recently proposed MFD algorithm[1]. The MFD algorithm uses the Spectral Graph Wavelet (SGW) transform in order to perform denoising directly in the spectral graph wavelet domain. Main function - Main_Demo provides an example of running our algorithm
