🧾 Medical Report OCR Parser 🧠 This project extracts structured JSON data from scanned medical reports using Tesseract OCR and Ollama LLMs (like Gemma or LLaMA). Ideal for automating data entry from printed lab reports.

🚀 Features 🖼️ OCR with Tesseract

📄 Medical text parsing using local LLMs via Ollama

🧠 Tested with models like gemma:2b, llama3.2:3b, and mistral

🧾 Outputs:

Clean text

Structured JSON

Debug logs

🔧 Setup

1. Clone this repo bash Copy Edit git clone https://github.com/Adars2005/ocr-hackathon cd ocr-hackathon
2. Install dependencies bash Copy Edit pip install -r requirements.txt
3. Install & Configure Tesseract Windows: Download Tesseract

Add path to tesseract.exe in your system PATH

4. Install Ollama and Pull Model bash Copy Edit ollama run llama3.2:3b

or try gemma2:b or mistral

▶️ Run the App bash Copy Edit python main.py --input "input_images/input.jpeg" 📁 Output ./output/text/ – Raw extracted text

./output/json/ – Final structured JSON