Web Application for Interpretation of Doctor's Handwritten Prescription and Suggesting the best Price Offer over Various e-Commerce Websites using AI

Abstract

Patient record-keeping is crucial for accurate diagnoses and treatment. In India, due to time constraints, most doctors manually write prescriptions, making it challenging for pharmacists to read them correctly. This increases the risk of dispensing the wrong medicine, which can have serious and even fatal consequences for patients. To address this problem, this research proposes an online handwritten medical prescription recognition system that allows users to scan prescriptions and compare prices across different websites. OCR techniques are used to recognize medicine names in handwritten prescriptions, while NLP techniques such as Cosine Similarity are employed to overcome the issue of misinterpretation. The review study focuses on Named Entity Recognition and Relation Extraction, which help identify named entities and extract relations between entities. Most state-of-the-art techniques are offline and computationally expensive, highlighting the need for understanding the deep-learning processes used in the proposed system. The system stores various features such as medicine names, uploaded images, user details, and search history to improve recognition accuracy. Additionally, new features are proposed to enhance accuracy further. Uses a local dataset from a pharmacy shop and past prescription records from local doctors.