XTD10

홈페이지 · 논문 4편

XTD10 is a dataset for cross-lingual image retrieval and tagging consisting of the MSCOCO2014 caption test dataset annotated in 7 languages that were collected using a crowdsourcing platform.