PROJECT TITLE:

Graph Signal Denoising via Trilateral Filter onGraph Spectral Domain - 2016

ABSTRACT:

This paper presents a graph signal denoising method with the trilateral filter defined within the graph spectral domain. The first trilateral filter (TF) is a data-dependent filter that's widely used as a foothold-preserving smoothing methodology for Image Processing. However, as a result of of the data-dependency, one cannot give its frequency domain representation. To overcome this downside, we tend to establish the graph spectral domain illustration of the info-dependent filter, i.e., a spectral graph TF (SGTF). This representation enables us to style an effective graph signal denoising filter with a Tikhonov regularization. Moreover, for the proposed graph denoising filter, we offer a parameter optimization technique to search for a regularization parameter that approximately minimizes the mean squared error w.r.t. the unknown graph signal of interest. Comprehensive experimental results validate our graph signal processing-based approach for images and graph signals.