BiyoVes - Python Library

AI-powered biometric, passport, and visa photo generation for Python.

Resources: PyPI · Source · Issues · License

Overview

BiyoVes provides a compact API for background removal, face alignment, standards-based cropping, and print-ready photo layouts.

Installation

pip install biyoves

Or from source:

git clone https://github.com/mehmetaytugyuruk/biyoves-python-library.git
cd biyoves-python-library
pip install -e .

Quick Start

Method 1: Class-Based Usage (Recommended)

from biyoves import BiyoVes

# Specify the photo path
img = BiyoVes("photo.jpg")

# Create a passport photo (2-up layout)
passport = img.create_image("vesikalik", "2li", "result_passport.jpg")

# Create a biometric photo (4-up layout)
biometric = img.create_image("biyometrik", "4lu", "result_biometric.jpg")

# US visa photo
us_visa = img.create_image("abd_vizesi", "2li", "result_us_visa.jpg")

# Schengen visa photo
schengen = img.create_image("schengen", "4lu", "result_schengen.jpg")

Method 2: Function-Based Usage

from biyoves import create_image

# Single-line processing
passport = create_image("photo.jpg", "vesikalik", "2li", "result.jpg")

Photo Types

• "biyometrik" - Standard biometric photo (50x60mm)
• "vesikalik" - Passport photo (45x60mm)
• "abd_vizesi" - US visa photo (50x50mm)
• "schengen" - Schengen visa photo (35x45mm)

Layout Types

• "2li" - 2 photos stacked vertically (2x1)
• "4lu" - 4 photos in a grid (2x2)

Features

• AI-powered automatic background removal
• Automatic face angle correction
• Automatic cropping to standard dimensions
• Print templates (2-up / 4-up layouts)
• Cut lines for print-ready output

Example Usage

from biyoves import BiyoVes

# Load a photo
img = BiyoVes("person.jpg")

# Save in different formats
img.create_image("vesikalik", "2li", "passport_2up.jpg")
img.create_image("vesikalik", "4lu", "passport_4up.jpg")
img.create_image("biyometrik", "2li", "biometric_2up.jpg")
img.create_image("abd_vizesi", "4lu", "us_visa_4up.jpg")

Requirements

• Python >= 3.7
• OpenCV
• NumPy
• ONNX Runtime

Models Used

This project uses the following ONNX models:

Model Directory: All models are stored in the src/biyoves/models/ directory.

Model Citations

• MODNet: Zhanghan Ke et al., "MODNet: Real-Time Trimap-Free Portrait Matting via Objective Decomposition," AAAI 2022.
• InsightFace: Jiankang Deng et al., "InsightFace: 2D and 3D Face Analysis Project."

Third-Party Models and Licensing

The BiyoVes source code is MIT-licensed. Bundled model weights retain their original terms and are not relicensed by this repository:

• MODNet code and published models are provided under Apache-2.0 by the MODNet project.
• InsightFace model-zoo weights, including the SCRFD and 2D106 components used here, are provided for non-commercial research purposes only according to the InsightFace model-zoo notice.

Review the upstream terms before redistributing the weights or using them in a commercial product.

License

The BiyoVes source code is released under the MIT License. Third-party model weights are governed by the terms listed above.