Search papers, labs, and topics across Lattice.
Huazhong University of Science and Technology
1
0
2
Calibrating text embeddings with visual prototypes boosts segmentation accuracy without any additional training, bridging the gap between generic concepts and specific visual instances.