Y. He, “Visual Semantics Allow for Textual Reasoning Better in Scene Text Recognition”, AAAI, vol. 36, no. 1, pp. 888-896, Jun. 2022.