From Pixels to Structure: Lightweight Vision-Language Models for Document OCR and Structured JSON Extraction.

Uddipan Basu Bir, Vincent Christlein, Andreas K. Maier, Mathias Zinnen

Browse the full ICDAR paper archive.