VisFocus: Prompt-Guided Vision Encoders for OCR-Free Dense Document Understanding

ECCV 2024

Project page