Fine-grained open-vocabulary object detection using NoctOWL
Demo of Talk2DINO, model presented at ICCV 2025.