Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
Nanjing University of Science and Technology, of Advanced Construction Machinery, State Key Laboratory of Intelligent Manufacturing
2
0
2
Semantic and attribute over-alignment in CD-FSS can be mitigated through a novel dual aggregation approach, leading to significant performance gains on novel classes.
Achieving state-of-the-art cross-domain segmentation performance without any training could revolutionize how we approach few-shot learning in diverse applications.