A scene-level annotated point cloud segmentation framework combining contrastive 2D-3D feature alignment with region-point consistency improves pseudo-label quality and achieves SOTA on ScanNet and S3DIS.
Condaformer: Disassembled transformer with local structure en- hancement for 3D point cloud understanding,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
High-quality Pseudo-labeling for Point Cloud Segmentation with Scene-level Annotation
A scene-level annotated point cloud segmentation framework combining contrastive 2D-3D feature alignment with region-point consistency improves pseudo-label quality and achieves SOTA on ScanNet and S3DIS.