Payment-Verification-OCR-2025

This repository contains a Python script for automating payment verification using Optical Character Recognition (OCR). The script efficiently extracts and validates key payment details from receipts or transaction records.

Features

• Extracts transaction IDs using PaddleOCR from payment screenshots.
• Supports popular platforms like PhonePe, Google Pay, and Paytm.
• Automatically matches extracted transaction IDs with backend transaction logs (Excel/CSV).
• Marks registrations as "verified" or "not verified".
• Generates a clean, downloadable Excel report of verified users.

Tech Stacks

• Python
• PaddleOCR
• Pandas
• Regex (for robust ID extraction)
• Google Colab / Jupyter compatible

Input Files

1. reg-data.csv – User registration data with screenshot links.
2. final transactions.xlsx – Drive log of successful transactions.
3. Second Transactions for ACM drive.xlsx – Additional transaction logs.

Output

• A downloadable Excel file containing:
  • Name
  • Extracted transaction ID
  • Status (Verified / Not Verified)

Supported Receipt Types

• PhonePe: UTR numbers starting with T followed by 21 digits.
• Google Pay: Transaction IDs like AXIS1234567890.
• Paytm: 12 to 15 digit numeric transaction reference numbers.

How It Works

1. Upload the registration Excel file.
2. The script:
   • Downloads each screenshot.
   • Applies OCR to extract transaction info.
   • Matches it with backend transaction logs.
3. Generates a final report with verification status.

Payment-Verification-OCR-2025/ ├── Transaction verification.ipynb ├── README.md └── sample_data/ ├── reg-data.csv ├── final transactions.xlsx └── Second Transactions for ACM drive.xlsx