Z Raisi, J Zelek - 2022 19th Conference on Robots and Vision (CRV), 2022 - computer.org
The performance of existing deep-learning scene text recognition-based methods fails
significantly on occluded text instances or even partially occluded characters in a text due to …