Fetching the paper…

DEVICE: Depth and Visual Concepts Aware Transformer for OCR-based Image Captioning · Around