Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition-' Washington , D . C . , June , 1983 OPTIMAL PERCEPTUAL INFERENCE

How researchers see this paper

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

PointNet: Deep Learning on Point Sets for 3D Classification and Segmentation (2016, background): “For example, rotating and translating points all together should not modify the global point cloud category nor the segmentation of the points.”
3DMatch: Learning Local Geometric Descriptors from RGB-D Reconstructions (2016, unknown): “3D ShapeNets [38] introduced 3D deep learning for modeling 3D shapes, and several recent works [23, 11, 33] also compute deep features from 3D data for the task of object retrieval and classification.”
Permutation-equivariant neural networks applied to dynamics prediction (2016, background): “That means that as long as one does a calculation that is applied identically to all members of the symmetry group, the output will be invariant to transformations applied to the input.”
Unsupervised Continual Learning Via Pseudo Labels (2021, methodology): “BIC [ Wu et al. , 2019 ] applies an additional linear model for bias correction after each incremental step using a small validation set containing balanced old and new class data.”
3DMatch: Learning the Matching of Local 3D Geometry in Range Scans (2016, background): “Moreover, compared to a 2D encoding of shape using depth patches, a voxelized 3D representation preserves real-world spatial scale information, is invariant to projective scaling, and naturally provides grounds for learning the invariance of spatial rotation.”