FOCUS masks everything outside a user-selected region with SAM, uses GPT-4o to propose an object list, and feeds that list to GroundingDINO or OWLv2 to output bounding boxes.
Yolo-world: Real-time open- vocabulary object detection
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CV 1years
2024 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
Fine-Grained Open-Vocabulary Object Recognition via User-Guided Segmentation
FOCUS masks everything outside a user-selected region with SAM, uses GPT-4o to propose an object list, and feeds that list to GroundingDINO or OWLv2 to output bounding boxes.