Improving OCR-Based Image Captioning by Incorporating Geometrical Relationship.

Jing Wang, Jinhui Tang, Mingkun Yang, Xiang Bai, Jiebo Luo

Browse the full CVPR paper archive.