Prescription Digitization & Validation Pipeline
Designed an end-to-end medical prescription digitization pipeline that converts handwritten doctor prescriptions into structured digital medical records. Combines Optical Character Recognition (OCR), NLP tokenization/stopword removal, text annotation, date validation (<30 days rule), string matching against licensed doctor directories, and intelligent medicine substitution backed by inventory knowledge bases.
Tech Stack
PythonFastAPIOCR (Tesseract/Vision)NLPString MatchingPostgreSQLKnowledge Base
Key Features
- Optical Character Recognition (OCR): High-accuracy text extraction from handwritten medical prescriptions.
- NLP Data Pre-Processing: Tokenization, stopword removal, and domain-specific medical entity extraction.
- Validation Logic: Automated verification of issuance dates (< 30 days) and doctor license verification.
- Medicine Substitution Engine: Intelligent mapping of prescribed brand names to available generic inventory substitutes.
Architecture & System Design