Dataset Description

• Homepage: COCO-CN homepage
• Paper: COCO-CN for Cross-Lingual Image Tagging, Captioning and Retrieval

Dataset Summary

COCO-CN is a bilingual image description dataset, developed based on (a subset of) MS-COCO with manually written Chinese sentences and tags. The dataset can be used for multiple tasks including image tagging, captioning and retrieval, all in a cross-lingual setting.

Languages

Chinese and English