aalquwayfili/ROCS

ROCS: Rare Objects in Cluttered Scenes Paper (ACML 2026) | Code A text-to-image retrieval benchmark for small, rare objects in cluttered scenes. Cluttered COCO and Flickr30K images are segmented with SAM3 to find single-instance, visually subordinate objects, then re-captioned with Qwen3-VL so each query names one such object. Retrieval is tested on whether the named rare object is found, not just the dominant scene. ROCS-COCO ROCS-Flickr30K Images 3,248 2,442… See the full description on the dataset page: https://huggingface.co/datasets/aalquwayfili/ROCS.

<iframe src="https://silytics.si/embed/item/525f008b-2e85-4913-b555-36ebde1f54b9" width="420" height="190" style="border:0" loading="lazy" title="Silytics entry"></iframe>

Silytics shows public metadata from Hugging Face and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.