1 paper
Shenghao Fu, Yukun Su, Fengyun Rao +3
Open-vocabulary object detection aims to detect arbitrary classes via text prompts. Methods without cross-modal fusion layers (non-fusion) offer faster inference by treating recogn…