Search papers, labs, and topics across Lattice.
Hikvision Research Institute
1
0
3
8
Visual prompts can now rival text prompts in open-vocabulary object detection, thanks to a novel framework that injects global discriminability.