Reproduce “Annotator-Free Object Detection via Foundation Model Collaboration” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Yasiru Ranasinghe, Vishal M. Patel
The authors turn unlabeled target-domain images into training data for an object detector by coordinating pretrained segmentation and vision-language models to propose boxes and labels, then training a detector to refine those proposals. On COCO val, the YOLO version reaches 60.8 AP versus 52.2 AP for T-Rex2 (Text), under exact category-name matching; this does not establish a win on every benchmark.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.