Handwritten medical prescriptions remain a common practice across healthcare systems, particularly in developing regions, yet they pose significant challenges due to poor legibility and inconsistent formats. Misinterpretation of these prescriptions by pharmacists or patients can result in serious consequences such as incorrect dosages, adverse drug reactions, and delayed treatments. To address this issue, this project proposes the design and development of an AI-Based Optical Character Recognition (OCR) System for Handwritten Medical Prescription Recognition and Interpretation. The system integrates deep learning models, specifically CNN-LSTM architectures, with Natural Language Processing (NLP) techniques to accurately extract key medical information such as patient and doctor names, prescribed drugs, dosages, and administration instructions. An image preprocessing pipeline including binarization, noise removal, and line segmentation enhances recognition accuracy, while integration with OCR engines like Tesseract ensures robust text detection. A web-based user interface, developed using Streamlit, enables users to upload scanned or photographed prescriptions and obtain structured, real-time outputs. The recognized data is securely stored in a database for easy retrieval and integration with pharmacy systems or electronic health records. Experimental validation highlights the system’s potential to significantly reduce human errors in prescription handling, improve workflow efficiency in healthcare settings, and contribute to digital healthcare transformation across multilingual and multicultural environments.
Sharjeel et al. (2025) studied this question.