Point Transformer

lastmod 2026-11-10created 2026-02-138 edits

Point Transformer is a deep learning model that applies the Transformer's self-attention to 3D point cloud processing. On irregular, unordered point sets, it computes attention over each point's neighboring points to extract features, and is used for classification and segmentation tasks. V2 and V3, which improve efficiency and scale, were released later.

See also

• Point Transformer V3