Intrinsic shape signatures: A shape descriptor for 3D object recognition

By Yu Zhong. 2009 IEEE 12th International Conference on Computer Vision Workshops, ICCV Workshops, 2009

829 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computational Mechanics, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/e37e9fe0917f972394ff4bc152caca1c83f5c35a

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 Computational Mechanics

In Computer Vision and Pattern Recognition

Citations in the corpus

Image Matching from Handcrafted to Deep Features: A Survey (2020, methodology): “Zhong (2009) introduced the intrinsic shape signature (ISS) method, in which saliency is derived from the eigenvalue decomposition of the scatter matrix of the support region.”
Review: deep learning on 3D point clouds (2020, background): “In the past few years, processing of point cloud for visual intelligence has been based on handcrafted features [1, 2, 3, 4, 5, 6].”
D3Feat: Joint Learning of Dense Detection and Description of 3D Local Features (2020, methodology): “Thus we compare our detector in repeatability with a learning-based method USIP [16] and hand-crafted methods: ISS [37], Harris-3D [13], and SIFT-3D [18] on the 3DMatch test set and KITTI test set.”
Recognizing Objects in 3D Point Clouds with Multi-Scale Local Features (2014, methodology): “For each model,Mi, nm keypoints are first detected by uniform sampling and then pruned by a resolution control strategy [4,21].”
Mid-Level Visual Representations Improve Generalization and Sample Efficiency for Learning Active Tasks (2018, background): “…set of approaches—e.g. fully supervised learning [23], self-supervised learning [32, 35, 7, 56, 55, 33], unsupervised learning [9, 4, 7, 56, 8, 41]—adopted to solve various standard perception tasks, e.g. depth estimation [24], object detection [23] and segmentation [43], pose estimation [57], etc.”
Autonomous robotic stone stacking with online next best object target pose planning (2017, methodology): “Besides the ISS descriptor we tested two additional descriptors, namely the Fast Point Feature Histogram (FPFH) descriptor [18] and the Rotational Projection Statistics (RoPS) descriptors [19].”
Learning to Detect Good 3D Keypoints (2017, methodology): “13, the relative ordering of hand-crafted detectors does change: on the one hand, the baseline uniform sampling approach performs better than KPQ, ISS, and Harris3D; on the other hand, NARF, whose performance was quite unsatisfactory on previous datasets, turns out the best hand-crafted detector onK”
Unique Signatures of Histograms for Local Surface Description (2010, methodology): “Intrinsic Shape Signatures [13] proposed an improvement of [12] based on a different partitioning of the 3D local volume as well as on a different definition of the local RF. Finally, Mian et al. [2] accumulate 3D histograms (Tensors) of mesh triangle areas within a cubic support.”
Performance Evaluation of 3D Keypoint Detectors (2012, methodology): “A notable scientific fervor has recently characterized the field of 3D detectors, leading to several relevant proposals (Mian et al. 2010; Chen and Bhanu 2007; Zhong 2009; Novatnack and Nishino 2008; Unnikrishnan and Hebert 2008; Akagunduz and Ulusoy 2007; Zaharescu et al. 2009; Fadaifard and Wolber”
3D Object Recognition in Cluttered Scenes with Local Surface Features: A Survey (2014, methodology): “Guo et al. [44] first constructed a unique, unambiguous and robust LRF by calculating the covariance matrix of all points lying on the local surface rather than just mesh vertices (in contrast to [3], [70]).”
SHOT: Unique signatures of histograms for surface and texture description (2014, methodology): “Intrinsic Shape Signatures [7] proposes an improvement of 3D Shape Context based on a different partitioning of the 3D local volume as well as on a different definition of the local RF. Point Feature Histograms (PFH) [27] and Fast Point Feature Histograms (FPFH) [16] accumulate in a 3D histogram thr”
Mid-Level Visual Representations Improve Generalization and Sample Efficiency for Learning Visuomotor Policies (2018, background): “depth estimation [18, 43], object classification [42, 32], detection [64, 25], segmentation [70, 30, 35], pose estimation [92, 9, 87], etc.”
3DFeat-Net: Weakly Supervised Local 3D Features for Point Cloud Registration (2018, methodology): “95 94.02% 7084 ISS [35] + CGF [15] 0 .”
Unique shape context for 3d data description (2010, background): “More specifically, in [22] 4 different local RFs have to be taken into account for each feature point to deal with the intrinsic ambiguity on the sign of the local RF unit vectors.”
Taskonomy: Disentangling Task Transfer Learning (2018, background): “They are often invariant to informative (but possibly distracting) features such as textures [50, 39, 25, 47, 16].”
Deep Learning Advances in Computer Vision with 3D Data (2017, background): “Indicative examples of ﬁxed-scale detectors are Local Surface Patches (LSPs) [Chen and Bhanu 2007], Intrinsic Shape Signatures (ISSs) [Zhong 2009], and the 3D detector proposed in Mian et al. [2010] termed as “KeyPoint Quality” (KPQ) in Tombari et al. [2013], while a popular adaptive-scale detector…”
Rotational Projection Statistics for 3D Local Surface Description and Object Recognition (2013, methodology): “Examples include spin image (Johnson and Hebert, 1999), Intrinsic Shape Signatures (ISS) (Zhong, 2009) and Mesh-HOG (Zaharescu et al., 2012).”
SpinNet: Learning a General Surface Descriptor for 3D Point Cloud Registration (2020, methodology): “*Equal contribution Early methods to extract local geometries include PS [9], 3DSC [19], ISS [67] and SHOT [40], which simply compute the low-level features such as faces [41, 64], corners [52], and handcrafted statistical histograms [43].”
A Comprehensive Performance Evaluation of 3D Local Feature Descriptors (2015, methodology): “Experimental results showed that intrinsic shape signatures (ISS) is the most effective 3D keypoint detector, while SHOT is amongst the best 3D local feature descriptors.”