Point Signatures: A New Representation for 3D Object Recognition

By C. Chua, R. Jarvis. 1997

How researchers see this paper

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

Citations in the corpus

VoxelNet: End-to-End Learning for Point Cloud Based 3D Object Detection (2017, methodology): “Some of the earlier methods for feature representation are [41, 8, 7, 19, 42, 35, 6, 27, 1, 36, 2, 25, 26].”
SHOT: Unique signatures of histograms for surface and texture description (2014, methodology): “In Point Signatures [21] the signature is given by the signed height of the 3D curve obtained by intersecting a sphere centered at the point with the surface.”
Automatic Reconstruction of As-Built Building Information Models from Laser-Scanned Point Clouds: A Review of Related Techniques | NIST (2010, background): “Other descriptors include point signatures [17], Euclidean 3D grid histograms [64], 3D shape contexts [28], and bitangent curves [98].”
On the Repeatability and Quality of Keypoints for Local Feature-based 3D Object Retrieval from Cluttered Scenes (2009, background): “Chua and Jarvis [7] introduced point signatures for 3D object recognition.”
Schelling points on 3D surface meshes (2012, background): “2002], points with unlikely local shape descriptors [Chua and Jarvis 1996; Johnson 2000], and points with local shape descriptors distinctive of an object class [Shilane and Funkhouser 2007].”
Intrinsic shape signatures: A shape descriptor for 3D object recognition (2009, methodology): “Subsequently, the methods of " point-signatures " [5] and " principle curvature " , proposed by Chua and Jarvis, were also used to match local surfaces of 3D objects.”
Fully automatic registration of multiple 3D data sets (2003, background): “Proposed encodings include spin-images [1], splashes [3], point signatures [4], harmonic shape images [5], spherical attribute images [6], and the tripod-operator [7].”
Recognizing Objects in Range Data Using Regional Point Descriptors (2004, background): “Methods which use regional point descriptors have proven successful in the context of image-based recognition [17][15][2] as well as 3D recognition and surface matching [22][13][5][21].”
A survey of content based 3D shape retrieval methods (2004, methodology): “Due to the complexity of their representation [ 28 , 58] these methods are very difficult to apply to 3D shape matching.”
Unique shape context for 3d data description (2010, methodology): “As for [5], the local RF relies on a Repeatable Axis (the normal of the feature point), then a full local RF is defined by adding the direction and sign of the global maximum of the signed height of the 3D curve obtained by intersecting a sphere centered in the point with the surface.”
Learning Object Bounding Boxes for 3D Instance Segmentation on Point Clouds (2019, methodology): “To extract features from 3D point clouds, traditional approaches usually craft features manually [4; 40].”
State of research in automatic as-built modelling (2015, methodology): “Other methods include free form surface registration [35] using angular invariant features [65], or modified ICP methods such as the dual interpolating point-to-surface method [121].”
Scale-hierarchical 3D object recognition in cluttered scenes (2010, background): “Chua and Jarvis [2] use the point signature which encodes the minimum distances of points on a 3D contour to a reference plane.”
A Survey on Human Motion Analysis from Depth Data (2013, methodology): “[90] use point signatures [91] and Gabor filters to detect some facial feature points from 3d and 2d data.”
A review of recent range image registration methods with accuracy evaluation (2007, methodology): “Point signature is a point descriptor introduced by Chua [9] and used to search for correspondences.”
An Efficient Multimodal 2D-3D Hybrid Approach to Automatic Face Recognition (2007, methodology): “Wang et al. [46] used Gabor Wavelet filters in the 2D domain and the point signatures [12] in the 3D domain and fused the results using SVM.”
Model globally, match locally: Efficient and robust 3D object recognition (2010, background): “Point Signatures [4] use the distance of neighbouring points to the normal plane of the reference point.”
2D and 3D face recognition: A survey (2007, background): “They find these ‘‘rigid’’ facial regions by using a Point Signature two-by-two comparison (Chua and Jarvis, 1997) among different facial expressions of the same person.”
Unique Signatures of Histograms for Local Surface Description (2010, methodology): “To prevent these shortcomings from invalidating the comparison we consider only the global maximum of the height [7] for PS and we add the sign disambiguation of [14] to EM (EM+SD), thereby obtaining two unique and unambiguous RFs.”