Introducing FastMRZ – Effortless MRZ Extraction Made Simple

Sivakumar Mahalingam — Tue, 31 Dec 2024 15:20:58 +0000

Overview

In today's world, where speed is everything, sensitive document processing has to be automated. Be it passport, visa, or ID card, extraction of MRZ data with accuracy and speed is very significant for developing reliable applications. This is where fastmrz comes into play as a lightweight and efficient Python package in making this MRZ data extraction from document images much easier.

GitHub Repository: FastMRZ Repo

PyPI: FastMRZ Python Package

What is FastMRZ?

FastMRZ is a Python package that uses computer vision techniques to extract MRZ information from images of documents. It provides an easy-to-use implementation that enables the extraction of MRZ, making it easier for developers who build systems for FinTech, security, and identity verification.

Why use FastMRZ?

1. Ease of Use: With a simple API, fastmrz gets you up and running in no time and does not require a steep learning curve.

2. Speed: Optimized for performance, it processes images fast enough to ensure the least latency in your applications.

3. Accuracy: It is designed to reliably detect and extract MRZ data, even from challenging image conditions.

4. Lightweight: It is built to keep dependencies at a minimum, hence easily integrable into any project.

5. Open Source: It is free and open-source, hence the community contributes and improves the package.

Also this python package,

Supports various document formats (passports, visas, ID cards).
Compatible with a wide range of image resolutions and formats.
Robust against skewed or distorted images.
Pythonic API for seamless integration.

Key Features

FastMRZ is unique in many ways with the combination of powerful features:

1. Intelligent MRZ Detection: The library applies sophisticated contour detection methods in order to find the exact position of the MRZ area within document images to accurately extract even if document orientation is not perfect.

2. Custom ONNX Models: We have integrated our very own, custom-trained ONNX models for improved recognition accuracy by availing the powers of machine learning while maintaining rapid speeds.

3. Built-in Validation: It features comprehensive checksum logic that will verify the extracted data, helping to catch any potential recognition errors or document tampering.

4. Flexible Output Formats: Whether it's structured JSON for easy integration with other systems or even raw text for simple processing, FastMRZ has you covered.

Under the Hood

FastMRZ is built on a very robust technical foundation, using several leading technologies:

OpenCV for image processing and MRZ region detection
Tesseract OCR for text recognition, with custom-trained data for MRZ-specific characters
NumPy for efficient numerical operations
ONNX to run optimized machine learning models.

Usage

Installation

You can install fastmrz via pip:

pip install fastmrz

How to Use fastmrz

Here’s a quick example to get started:

from fastmrz import FastMRZ
import json

fast_mrz = FastMRZ()

passport_mrz = fast_mrz.get_mrz("../data/passport_uk.jpg")
print("JSON:")
print(json.dumps(passport_mrz, indent=4))

print("\n")

passport_mrz = fast_mrz.get_mrz("../data/passport_uk.jpg", raw=True)
print("TEXT:")
print(passport_mrz)

OUTPUT:

JSON:
{
    "mrz_type": "TD3",
    "document_type": "P",
    "country_code": "GBR",
    "surname": "PUDARSAN",
    "given_name": "HENERT",
    "document_number": "707797979",
    "nationality": "GBR",
    "date_of_birth": "1995-05-20",
    "sex": "M",
    "date_of_expiry": "2017-04-22",
    "status": "SUCCESS"
}


TEXT:
P<GBRPUDARSAN<<HENERT<<<<<<<<<<<<<<<<<<<<<<<
7077979792GBR9505209M1704224<<<<<<<<<<<<<<00

Understanding MRZ Types

FastMRZ supports various MRZ formats, including:

TD-1 format (3 lines, 30 characters each) commonly used in ID cards
TD-2 format (2 lines, 36 characters each) found in some visas and older IDs
TD-3 format (2 lines, 44 characters each) used in international passports

The library automatically detects the format and processes it accordingly, making it versatile for different document types.

Applications

Identity Verification: Automate ID validation processes in security and travel.
FinTech: Enhance KYC processes with seamless document verification.
Travel and Immigration: Streamline data capture for passports and visas.
Government Applications: Assist in border control and other regulatory activities.

Conclusion

FastMRZ enables more accessibility and efficiency of document processing. Be it the implementation of a border control system, working on identity verification software, or any document processing application-the MRZ extraction definitely will be done with confidence by using the power of FastMRZ.

You can fork the project on GitHub and star it, if you find it useful. Your stars help drive the continued development of this open-source tool.

DEV Community: Sivakumar Mahalingam