Unique shape context for 3d data description

By Federico Tombari, Samuele Salti, Luigi Di Stefano. Eurographics Workshop on 3D Object Retrieval, 2010

377 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition, Geology

View Paper: https://www.semanticscholar.org/paper/2ba27db4559b1478c5aaed6727ae79d438be380f

How researchers see this paper

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

In Aerospace Engineering

In Computer Vision and Pattern Recognition

In Geology

Citations in the corpus

Deep Global Registration (2020, background): “Traditional hand-crafted features commonly summarize pairwise or higher-order relationships in histograms [20, 37, 40, 36, 35].”
CoFiNet: Reliable Coarse-to-fine Correspondences for Robust Point Cloud Registration (2021, methodology): “From early-stage hand-crafted methods [1, 2, 3, 4, 5] to recently emerged deep learning-based approaches [6, 7, 8, 9, 10, 11, 12], many works contributed to improving the reliability of correspondences.”
Performance of global descriptors for velodyne-based urban object recognition (2014, background): “[10] proposed Unique 3D SC on the Stanford dataset which employs a unique, unambiguous local Reference Frame (RF).”
An Overview to Visual Odometry and Visual SLAM: Applications to Mobile Robotics (2015, methodology): “Some of the most well-known 3D Descrip-tor methods are Point Feature Histogram (PFH) [98], Fast Point Feature Histogram (FPFH) [99], 3D Shape Context (3DSC) [46], Unique Shape Context (USC) [112] and Sig-natures of Histograms of Orientations (SHOT) [112].”
PREDATOR: Registration of 3D Point Clouds with Low Overlap (2020, background): “Local 3D feature descriptors: Early local descriptors for point clouds [19, 31, 30, 37, 36] aimed to characterise the local geometry by using hand-crafted features.”
Tangent Convolutions for Dense Prediction in 3D (2018, background): “(The same assumption concerning normal vector estimation is made by both classic and contemporary geometric feature descriptors [20, 12, 45, 53, 46, 23].”
Fully Convolutional Geometric Features (2019, methodology): “Existing learning-based features rely on low-level geometric characteristics as input: e.g., angular deviation [7, 6, 26, 29, 23], point distributions [17, 16, 21], or volumetric distance functions [36, 11].”
Intensity Scan Context: Coding Intensity and Geometry Relations for Loop Closure Detection (2020, background): “Guo et al. justify that intensity information can be distinctive for places and propose a novel local descriptor called intensity signature of histograms of orientations (ISHOT) that consists of both geometry and intensity information.”
RPM-Net: Robust Point Matching Using Learned Features (2020, methodology): “Generally, these descriptors accumulate measurements (typically number of points) into histograms according to their spatial coordinates [9, 14, 32], or their geometric attributes such as curvature [4] or normals [28].”
PPFNet: Global Context Aware Local Features for Robust 3D Point Matching (2018, methodology): “The associated local reference frame [41] then aligns the patches with canonical axes.”
Pointwise Rotation-Invariant Network with Adaptive Sampling and 3D Spherical Voxel Convolution (2018, methodology): “ted to manual designed features, including Point Feature Histograms (PFH)[23], Fast Point Feature Histograms (FPFH)[22], Signature of Histogram Orientations (SHOT)[24], and Unique Shape Contexts (USC)[26]. These descriptors rely on delicate hand-craft design, and could only capture low-level geometr”
The Perfect Match: 3D Point Cloud Matching With Smoothed Densities (2018, methodology): “Many approaches including SHOT [39], RoPS [12], USC [38] and TOLDI [43] try to ﬁrst estimate a unique local reference frame (LRF), which is typically based on the eigenvalue decomposition of the sample covariance matrix of the points in neighborhood of the interest point.”
3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration (2018, methodology): “Other 3D descriptors [13,30] build histograms based on the number of points falling into each spatial bin around the keypoint.”
Image Matching from Handcrafted to Deep Features: A Survey (2020, background): “These histograms include local surface patch (Chen and Bhanu 2007), THRIFT (Flint et al. 2007), point feature histogram (Rusu et al. 2008), fast point feature histogram (Rusu et al. 2009) and signature of histogram of orientations (Tombari et al. 2010b).”
3D Object Recognition in Cluttered Scenes with Local Surface Features: A Survey (2014, neutral): “[114] proposed a unique shape context (USC)”
A comprehensive survey on point cloud registration (2021, neutral): “2 m USC [93] 43 rmseP( 0 .”