This document summarizes the steps and development process of implementing Thai OCR extraction and management using Tesseract.js, Google Vision, MongoDB, and React.
- Started with Tesseract.js for OCR extraction.
- Explored Google Vision for improved performance.
- Tested the speed of Google Vision on the local computer.
- Parsed extracted data using the parse function.
- Faced challenges due to high variations in data.
- Adopted Google's newly launched Gemini-Pro for improved data extraction in JSON format.
- Utilized MongoDB Compass for initial database management.
- Migrated to MongoDB Atlas for better scalability and management.
- Implemented CRUD operations for OCR records in the MongoDB database.
-
Add OCR Record:
POST - https://qoala-ocr-backend-production.up.railway.app/api/add-to-database
-
Retrieve OCR Data:
GET - https://qoala-ocr-backend-production.up.railway.app/api/ocr-records
-
Update OCR Record:
PUT - https://qoala-ocr-backend-production.up.railway.app/api/ocr-record/:id
-
Delete OCR Record:
DELETE - https://qoala-ocr-backend-production.up.railway.app/api/ocr-record/:id
- Created a React component for displaying OCR data.
- Implemented a UI form for updating and deleting OCR records.
- Added functionality to filter and sort displayed data.
-
Upload Thai ID: Upon uploading a Thai ID image, receive a neatly formatted JSON code, allowing users to copy and use the data effortlessly.
-
View All Data: Explore a comprehensive display of Thai OCR data in a clean and organized tabular format.
-
Filter Data: Easily refine data by Date of Birth (DOB) or other criteria, providing quick access to specific records.
-
Update User Data: Effortlessly modify OCR records by entering the Record ID and selecting the desired field for an update.
-
Delete User Entry: Delete entries seamlessly using the Record ID, streamlining the removal process.
- Successfully integrated OCR extraction, database management, and frontend UI.
- Achieved robust CRUD operations on the MongoDB Atlas database.