Abstract: Learning a powerful representation from point clouds is a fundamental and challenging problem in the field of computer vision. Different from images where RGB pixels are stored in the regular grid, for point clouds, the underlying semantic and structural information of point clouds is the spatial layout of the points. Moreover, the properties of challenging in-context and background noise pose more challenges to point cloud analysis. One assumption is that the poor performance of the classification model can be attributed to the indistinguishable embedding feature that impedes the search for the optimal classifier. This work offers a new strategy for learning powerful representations via a contrastive learning approach that can be embedded into any point cloud classification network. First, we propose a supervised contrastive classification method to implement embedding feature distribution refinement by improving the intra-class compactness and inter-class separability. Second, to solve the confusion problem caused by small inter-class compactness and inter-class separability. Second, to solve the confusion problem caused by small inter-class variations between some similar-looking categories, we propose a confusion-prone class mining strategy to alleviate the confusion effect. Finally, considering that outliers of the sample clusters in the embedding space may cause performance degradation, we design an entropy-aware attention module with information entropy theory to identify the outlier cases and the unstable samples by measuring the uncertainty of predicted probability. The results of extensive experiments demonstrate that our method outperforms the state-of-the-art approaches by achieving 82.9% accuracy on the real-world ScanObjectNN dataset and substantial performance gains up to 2.9% in DCGNN, 3.1% in PointNet++, and 2.4% in GBNet.

Attention-Based Improvement on 3D Point Cloud Classification

3D Point Cloud Classification Method Based on Multiple Attention Mechanism and Dynamic Graph Convolution

Investigating Attention Mechanism in 3D Point Cloud Object Detection

Multi-Head Self-Attention for 3D Point Cloud Classification

Attention Models for Point Clouds in Deep Learning: A Survey

A 3D Point Cloud Classification Method Based on Adaptive Graph Convolution and Global Attention

Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification

PointCloud-At: Point Cloud Convolutional Neural Networks with Attention for 3D Data Processing

DSACNN: Dynamically Local Self-Attention CNN for 3D Point Cloud Analysis.

Linked Attention-Based Dynamic Graph Convolution Module for Point Cloud Classification

Attentional Keypoint Detection on Point Clouds for 3D Object Part Segmentation

Dual-Graph Attention Convolution Network for 3-D Point Cloud Classification

A feature fusion-based attention graph convolutional network for 3D classification and segmentation

A Point Cloud Classification Method Based on Channel and Spatial Attention Mechanism

Point-based Attention Convolutional Neural Networks for Point Clouds Semantic Segmentation

Improving performance of deep learning models for 3D point cloud semantic segmentation via attention mechanisms

CloudAttention: Efficient Multi-Scale Attention Scheme For 3D Point Cloud Learning

Rethinking Attention Module Design for Point Cloud Analysis

Point Attention Network for Semantic Segmentation of 3D Point Clouds

Contrastive Embedding Distribution Refinement and Entropy-Aware Attention for 3D Point Cloud Classification

3D Point Cloud Classification and Segmentation Based on Dual Attention and Weighted Dynamic Graph Convolution