Improving OCR using internal document redundancy

Diego Belzarena, Seginus Mowlavi, Aitor Artola +9

Current OCR systems are based on deep learning models trained on large amounts of data. Although they have shown some ability to generalize to unseen data, especially in detection…