novel classes

5 papers

• DAA: Amplifying Unknown Discrepancy for Test-Time Discovery
• DON’T NEED RETRAINING: A Mixture of DETR and Vision Foundation Models for Cross-Domain Few-Shot Object Detection
• FedLPA: Local Prior Alignment for Heterogeneous Federated Generalized Category Discovery
• Novel Class Discovery for Point Cloud Segmentation via Joint Learning of Causal Representation and Reasoning
• Weak-shot Keypoint Estimation via Keyness and Correspondence Transfer