Point Cloud Classification Using Content-Based Transformer via Clustering in Feature SpaceOACSTPCD

Recently,there have been some attempts of Trans-former in 3D point cloud classification.In order to reduce com-putations,most existing methods focus on local spatial attention,but ignore their content and fail to establish relationships between distant but relevant points.To overcome the limitation of local spatial attention,we propose a point content-based Transformer architecture,called PointConT for short.It exploits the locality of points in the feature space(content-based),which clusters the sampled points with similar features into the same class and com-putes the self-attention within each class,thus enabling an effec-tive trade-off between capturing long-range dependencies and computational complexity.We further introduce an inception fea-ture aggregator for point cloud classification,which uses parallel structures to aggregate high-frequency and low-frequency infor-mation in each branch separately.Extensive experiments show that our PointConT model achieves a remarkable performance on point cloud shape classification.Especially,our method exhibits 90.3％Top-1 accuracy on the hardest setting of ScanOb-jectNN.Source code of this paper is available at https://github.com/yahuiliu99/PointConT.