Estimating surface normals in noisy point cloud data

By Niloy J. Mitra, A. Nguyen, Leonidas J. Guibas. 2003

748 citations recorded by the source corpus.

Subfields: Computational Mechanics, Computer Graphics and Computer-Aided Design, Environmental Engineering, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computational Mechanics

In Computer Graphics and Computer-Aided Design

In Environmental Engineering

In 3D Vision & Pose

Citations in the corpus

PointNet++: Deep Hierarchical Feature Learning on Point Sets in a Metric Space (2017, methodology): “Previously several approaches have been developed regarding this [19; 17; 2; 6; 7; 30] either in geometry processing community or photogrammetry and remote sensing community.”
DIMENSIONALITY BASED SCALE SELECTION IN 3D LIDAR POINT CLOUDS (2012, background): “Starting for the minimal possible subset around the point of interest, the neighborhood is iteratively increased until the angular variance reaches a predefined threshold (Mitra et al., 2004).”
Unique Signatures of Histograms for Local Surface Description (2010, methodology): “It is worth pointing out that, compared to [15] and [16], in our proposal the third eigenvector no longer represents the TLS estimation of the normal direction and sometimes it notably differs from it.”
Just Look at the Image: Viewpoint-Specific Surface Normal Prediction for Improved Multi-View Reconstruction (2016, background): “If normal vectors are required , they are found in post-processing by fitting local tangent planes to the point cloud [21] [29].”
Geometric Back-Projection Network for Point Cloud Classification (2019, background): “he main drawback is the lack of geometric information compared with well-constructed mesh or volumetric data. In order to acquire more underlying knowledge of point clouds, conventional methods [18], [42], [43] attempted to estimate the geometry of point clouds i.e., face, normals, curvature, etc. I”
Segmentation and Recognition Using Structure from Motion Point Clouds (2008, background): “These are intentionally weak; stronger features would not work with the sparse noisy point clouds, though dense feature tracking could someday enable one to apply [25].”
A Theoretical and Computational Framework for Isometry Invariant Recognition of Point Cloud Data (2005, background): “The importance of this type of shape representation is leading to an increase in the fundamental study of point clouds, from its basic geometric characteristics such as curvature [43] and intrinsic distances and geodesics [42], to intrinsic dimensionality and topological structure [2, 13, 16, 17, 22”
Persistent Point Feature Histograms for 3D Point Clouds (2008, background): “Unfortunately, some of the above mentioned methods for computing an optimal scale require additional thresholds, such as d 1 and d 2 which are determined empirically in [8], and estimated using linear least-squares in [9] when knowledge about ground truth normal exists.”
SHOT: Unique signatures of histograms for surface and texture description (2014, methodology): “A Total Least Squares (TLS) estimation of the normal direction is obtained in Hoppe’s and Mitra’s works [32], [33] by EigenValue Decomposition (EVD) of the covariance matrix, M, of the k−nearest neighbors pi of the point:”
Accurate 3D comparison of complex topography with terrestrial laser scanner: Application to the Rangitikei canyon (N-Z) (2013, methodology): “This highlights the intuitive notion that the scale at which the normal should be estimated depends on the surface roughness (Mitra and Nguyen, 2003; Lalonde et al., 2005; Bae et al., 2009).”
Signing the Unsigned: Robust Surface Reconstruction from Raw Pointsets (2010, neutral): “…ﬁnding and orienting normals from raw geometric data can be as hard as reconstructing the whole surface itself: while there are several options to reliably estimate normal directions [MNG04,ACSTD07], robust normal orientation is considerable harder as recently reminded in Huang et al. [HLZ ∗ 09].”