A Curated Bangladesh-Based Dataset of Handwritten and Printed Prescription Images

This dataset comprises 200 de-identified prescription images collected in Bangladesh, including printed and handwritten prescriptions written in Bangla, English, and mixed Bangla-English formats. Direct identifiers such as patient and physician names, phone numbers, physician degrees, signatures, chamber details, and registration numbers have been removed for privacy protection. The dataset is distributed as a ZIP archive containing an images folder with the prescription images, a labels folder with corresponding text files containing YOLO-style bounding box coordinates for medicine-name regions, and a CSV file listing the image number and the number of annotated medicine boxes for each prescription image.

The dataset is intended for research and educational use in optical character recognition, handwriting recognition, prescription parsing, multilingual medical document understanding, object detection, and healthcare document analysis.