Textural or Textual: How Vision-Language Models Read Text in Images.

Hanzhang Wang, Qingyuan Ma

Browse the full ICML paper archive.