Optical Character Recognition.

This is a full stack OCR application to retreive JSON format data from the Thai ID cards.

Dependency.

Makes use of the Google Cloud Vision Api for retreiving text from the images.

Tech Stack.

1. Reactjs.
2. Mongoose-- Mongodb object modelling.
3. Mongodb as database.
4. Express server.
5. TailwindCss styling.

Description.

The application takes an image from the user, as an input. The image must follow the format, similar to that of a Thai Id card, images other than that may give unpredictable outputs. Results are shown in the JSON format, along with the number of success/failure operations happening.

Setting up the project.

1. cd frontend
2. run the command npm install
3. In the frontend folder add .env file in the following format---

VITE_PRODUCT_ID = <your_google_cloud_project_id_comes_here>

5. In the frontend folder, src/utils go to origin.js change origin to http://localhost:3000
6. run the command npm run dev
7. cd backend
8. run the command npm install
9. In the backend folder add a secret file of the json format obtained from google cloud vision api, which contains secret public/private key in json.
10. In the backend folder add .env file in the following format---

PROJECT_ID = <your_google_cloud_project_id_comes_here>
SECRET_KEY_PATH = <secret_file_name>.json
uri =  <mongodb_uri_connect_comes_here>

11. run the command npm run dev

Live Working

https://ocr-nine-sable.vercel.app/

note- The backend is deployed on Free tier of Render, so may take time to respond, kindly wait.