MS COCO

MS-COCO is a dataset that contains simple scenes used to evaluate the performance of vision-language models in visual reasoning tasks.

Papers using MS COCO (3)

Do All Visual Tokens Matter Equally? Object-Evidence Preserving Token Merging for Vision-Language Retrieval2026Evolution of Accuracy and Visual-Cognitive Errors in a Decade of Vision-Language AI Models2026EmCom-Diffusion: Probing Visual Reflection in Emergent Languages via Image Generation2026