Mesh Denoising via Adaptive Consistent Neighborhood.

Mingqiang Guo,Zheng Liu,Chengde Han,Saishang Zhong,Zhenzhen Song,Ruina Lv

doi:10.3390/s21020412

Mingqiang Guo, Zheng Liu + Show 4 more

Open Access

https://doi.org/10.3390/s21020412

Copy DOI

Abstract

In this paper, we propose a novel guided normal filtering followed by vertex updating for mesh denoising. We introduce a two-stage scheme to construct adaptive consistent neighborhoods for guided normal filtering. In the first stage, we newly design a consistency measurement to select a coarse consistent neighborhood for each face in a patch-shift manner. In this step, the selected consistent neighborhoods may still contain some features. Then, a graph-cut based scheme is iteratively performed for constructing different adaptive neighborhoods to match the corresponding local shapes of the mesh. The constructed local neighborhoods in this step, known as the adaptive consistent neighborhoods, can avoid containing any geometric features. By using the constructed adaptive consistent neighborhoods, we compute a more accurate guide normal field to match the underlying surface, which will improve the results of the guide normal filtering. With the help of the adaptive consistent neighborhoods, our guided normal filtering can preserve geometric features well, and is robust against complex shapes of surfaces. Intensive experiments on various meshes show the superiority of our method visually and quantitatively.

Highlights

A triangulated mesh is one of the typical data types for representing 3D models.Commonly, triangulated meshes can be generated by using the original 3D coordinate datas that collected by 3D model scanning equipments, such as Kinect, laser scanner, CT, etc
Sensors 2021, 21, 412 structures, multi-scale features, and fine details). In view of these issues, we propose a guided normal filtering based on adaptive consistent neighborhoods for mesh denoising
Experiments demonstrate that our method outperforms the existing state-of-the-art mesh denoising methods qualitatively and quantitatively

Summary

Introduction

A triangulated mesh is one of the typical data types for representing 3D models.Commonly, triangulated meshes can be generated by using the original 3D coordinate datas that collected by 3D model scanning equipments, such as Kinect, laser scanner, CT, etc. There are many noises in the original 3D coordinate datas, and the noises are generated in the 3D models reconstruction process [1,2]. The classical isotropic methods [6,7] mainly focus on removing the surface noise, but they neglect to preserve geometric features during the filtering process. These isotropic methods tend to produce denoised results with significant shape distortion. Guided normal filtering [21] followed by vertex updating is a well developed featurepreserving mesh denoising framework. The key of guided normal filtering is that it provides a robust guidance normal for each face of the mesh. According to the Equation (1), the filtered face normals of the surface can be obtained, we reconstruct vertex positions to match these filtered normals

Methods

Findings

Discussion

Conclusion