Geometric and Learning-based Mesh Denoising: A Comprehensive Survey

Mesh denoising is a fundamental problem in digital geometry processing. It seeks to remove surface noise while preserving intrinsic signals as accurately possible. While traditional wisdom has been built upon specialized priors smooth surfaces, learning-based approaches are making their debut with great success generalization and automation. In this work, we provide comprehensive review of the advances mesh denoising, containing both geometric recent methods. First, familiarize readers tasks, summarize four common issues denoising. We then two categorizations existing Furthermore, three important categories, including optimization-, filter-, data-driven-based techniques, introduced analyzed detail, respectively. Both qualitative quantitative comparisons illustrated, demonstrate effectiveness state-of-the-art Finally, potential directions future work pointed out solve problems these approaches. A benchmark also researchers will easily conveniently evaluate methods To aid reproducibility, release our datasets used results at https://github.com/chenhonghua/Mesh-Denoiser.