Mesh Denoising Using Quadric Error Metric

In this work we present a new method for mesh denoising that uses an operator based on the Quadric Error Metric. This operator is able to estimate the local shape of the surface for each vertex, despite severe noise condition, distinguishing corners, edges and smooth regions in order to best adjust the vertex geometry to recover piecewise smoothing while preserving sharp features. Our method results in a simple algorithm for mesh denoising that can also be used to enhance sharp features present in the surface corrupted by noise. A frequency response analysis is also presented in order to evaluate the characteristics of this operator in the frequency spectrum of the mesh.