Conference
Aug 2026
SemanticAdapter: parameter-efficient adaptation for label-efficient 3D point cloud semantic segmentation
SemanticAdapter is proposed, a parameter-efficient adaptation method that freezes the pretrained 3D encoder and introduces two lightweight components: a Cross-Modal Attention (CMA) module using point cloud features as queries and CLIP text embeddings as keys/values for explicit semantic alignment.
Yu Zhang, Yulin Hou
· International Conference on... · 0 citations