LIMA, L. S.; http://lattes.cnpq.br/2986992748216346; LIMA, Leandro de Souto.
Résumé:
Date recognition in medical prescriptions is a critical task to ensure patient safety and the efficiency of medical processes. This study proposes a comparative analysis between optical character recognition (OCR) models and handwritten text recognition models to identify and extract dates from medical prescriptions. Accurate date recognition is essential to avoid medication errors and ensure proper treatment administration. We evaluated the accuracy, speed, and robustness of the models in different scenarios and handwriting types, considering common variations found in medical prescriptions. The results show that OCR models have advantages in terms of speed and accuracy in printed texts, while handwritten text recognition models excel in interpreting varied handwritings. These findings provide valuable insights for the development of more effective date recognition systems in clinical settings.
