DNF-Net: A Deep Normal Filtering Network for Mesh Denoising

This paper presents a deep normal filtering network, called DNF-Net, for mesh denoising. To better capture local geometry, our network processes the in terms of patches extracted from mesh. Overall, DNF-Net is an end-to-end that takes facet normals as inputs and directly outputs corresponding denoised patches. In this way, we can reconstruct geometry with feature preservation. Besides overall architecture, contributions include novel multi-scale embedding unit, residual learning strategy to remove noise, deeply-supervised joint loss function. Compared recent data-driven works on denoising, does not require manual input extract features utilizes training data enhance its denoising performance. Finally, present comprehensive experiments evaluate method demonstrate superiority over state art both synthetic real-scanned meshes.