BishtRef
Articles Journals Publishers Authors Subjects Most Cited New Papers Year Stats Open Access
journal article Jun 25, 2019

PointCleanNet: Learning to Denoise and Remove Outliers from Dense Point Clouds

Marie‐Julie Rakotosaona Vittorio La Barbera Paul Guerrero ORCID Niloy J. Mitra ORCID Maks Ovsjanikov
Computer Graphics Forum Vol. 39 No. 1 pp. 185-203 · Wiley
View at Publisher Save 10.1111/cgf.13753
Abstract
AbstractPoint clouds obtained with 3D scanners or by image‐based reconstruction techniques are often corrupted with significant amount of noise and outliers. Traditional methods for point cloud denoising largely rely on local surface fitting (e.g. jets or MLS surfaces), local or non‐local averaging or on statistical assumptions about the underlying noise model. In contrast, we develop a simple data‐driven method for removing outliers and reducing noise in unordered point clouds. We base our approach on a deep learning architecture adapted from PCPNet, which was recently proposed for estimating local 3D shape properties in point clouds. Our method first classifies and discards outlier samples, and then estimates correction vectors that project noisy points onto the original clean surfaces. The approach is efficient and robust to varying amounts of noise and outliers, while being able to handle large densely sampled point clouds. In our extensive evaluation, both on synthetic and real data, we show an increased robustness to strong noise levels compared to various state‐of‐the‐art methods, enabling accurate surface reconstruction from extremely noisy real data obtained by range scans. Finally, the simplicity and universality of our approach makes it very easy to integrate in any existing geometry processing pipeline. Both the code and pre‐trained networks can be found on the project page (https://github.com/mrakotosaon/pointcleannet).
Topics

No keywords indexed for this article. Browse by subject →

References
60
[1]
Computing and rendering point set surfaces

M. Alexa, J. Behr, D. Cohen-Or et al.

IEEE Transactions on Visualization and Computer Gr... 10.1109/tvcg.2003.1175093
[2]
Achlioptas P. (2018)
[5]
Geometric Deep Learning: Going beyond Euclidean data

Michael M. Bronstein, Joan Bruna, Yann LeCun et al.

IEEE Signal Processing Magazine 10.1109/msp.2017.2693418
[7]
A Non-Local Algorithm for Image Denoising

A. Buades, B. Coll, J.-M. Morel

2005 IEEE Computer Society Conference on Computer... 10.1109/cvpr.2005.38
[9]
Barrow H. G. (1977)
[10]
A Survey of Surface Reconstruction from Point Clouds

Matthew Berger, Andrea Tagliasacchi, Lee M. Seversky et al.

Computer Graphics Forum 10.1111/cgf.12802
[11]
Chambolle A. "An introduction to total variation for image analysis" Theoretical Foundations and Numerical Methods for Sparse Recovery (2010)
[12]
Chazal F. "Geometric inference for probability measures" FOCS (2011)
[14]
Cazals F. "Estimating differential quantities using polynomial fitting of osculating jets" CAGD 22:2 (2005)
[15]
Cazals F. (2007)
[18]
Deschaud J.‐E. "Point cloud non local denoising using local surface descriptor similarity" IAPRS (2010)
[29]
Gross M. (2011)
[30]
A review of algorithms for filtering the 3D point cloud

Xian-Feng Han, Jesse S. Jin, Ming-Jie Wang et al.

Signal Processing: Image Communication 10.1016/j.image.2017.05.009
[32]
He K. "Delving deep into rectifiers: Surpassing human‐level performance on imagenet classification" ICCV (2015)
[33]
Deep Convolutional Neural Network for Inverse Problems in Imaging

Kyong Hwan Jin, Michael T. McCann, Emmanuel Froustey et al.

IEEE Transactions on Image Processing 10.1109/tip.2017.2713099
[34]
Jaderberg M. "Spatial transformer networks" NIPS (2015)
[35]
Screened poisson surface reconstruction

Michael Kazhdan, Hugues Hoppe

ACM Transactions on Graphics 10.1145/2487228.2487237
[38]
Mairal J. (2010)
[39]
Masci J. "Geodesic convolutional neural networks on Riemannian manifolds" ICCV Workshops (2015)
[40]
Point Cloud Denoising via Moving RPCA

E. Mattei, A. Castrodad

Computer Graphics Forum 10.1111/cgf.13068
[41]
Mederos B. (2003)
[44]
Qi C. R. "Pointnet: Deep learning on point sets for 3d classification and segmentation" Computer Vision and Pattern Recognition (2017)
[45]
Qi C. R. "Pointnet++: Deep hierarchical feature learning on point sets in a metric space" NIPS (2017)
[50]
Wei L. "Dense human body correspondences using convolutional networks" Proceedings of the Conference on Computer Vision and Pattern Recognition (CVPR) (2016)

Showing 50 of 60 references

Metrics
282
Citations
60
References
Details
Published
Jun 25, 2019
Vol/Issue
39(1)
Pages
185-203
License
View
Authors
M
Marie‐Julie Rakotosaona

LIX École Polytechnique, CNRS

V
Vittorio La Barbera

University College London

P
Paul Guerrero

University College London

N
Niloy J. Mitra

University College London

M
Maks Ovsjanikov

LIX École Polytechnique, CNRS

Cite This Article
Marie‐Julie Rakotosaona, Vittorio La Barbera, Paul Guerrero, et al. (2019). PointCleanNet: Learning to Denoise and Remove Outliers from Dense Point Clouds. Computer Graphics Forum, 39(1), 185-203. https://doi.org/10.1111/cgf.13753
Related

You May Also Like

Efficient RANSAC for Point‐Cloud Shape Detection

R. Schnabel, R. Wahl · 2007

1,591 citations

On Visual Similarity Based 3D Model Retrieval

Ding‐Yun Chen, Xiao‐Pei Tian · 2003

1,045 citations

Super 4PCS Fast Global Pointcloud Registration via Smart Indexing

Nicolas Mellado, Dror Aiger · 2014

513 citations

Force‐Directed Edge Bundling for Graph Visualization

Danny Holten, Jarke J. van Wijk · 2009

406 citations

A Survey of Surface Reconstruction from Point Clouds

Matthew Berger, Andrea Tagliasacchi · 2016

395 citations