Skip to content

Latest commit

 

History

History
67 lines (49 loc) · 2.68 KB

README.md

File metadata and controls

67 lines (49 loc) · 2.68 KB
Raw

NOTICE

This software (or technical data) was produced for the U. S. Government under contract, and is subject to the Rights in Data-General Clause 52.227-14, Alt. IV (DEC 2007)

(C) 2019 The MITRE Corporation. All Rights Reserved. Approved for Public Release; Distribution Unlimited. Public Release Case Number 18-0812.

Overview

The Biometric Image Quality Toolkit (BIQT) is an extensible framework for evaluating the quality of biometric images. It exposes a vendor-agnostic interface through which new modalities and algorithms can be integrated into the framework, and it also exposes a high-level Quality API which can be used to integrate the framework into larger systems.

The BIQT Architecture

Quick Start

Use the ghcr.io/mitre/biqt image hosted on GitHub to access a preconfigured Ubuntu 22.04 environment with the BIQT command-line interface. This image includes the BIQTFace, BIQTIris, and BIQTContactDetector providers.

The following Linux-based example shows how to start a mitre/biqt container capable of accessing and using the biometric images stored in the directory /path/to/biometric/images on the host filesystem.

user@localhost:~/$ docker run --rm -v /path/to/images:/data -it ghcr.io/mitre/biqt
[root@66d3679381d0 /]# biqt -m iris iris-image-001.png

Installation

The framework targets Ubuntu 22.04.

Build and install instructions for Docker and Ubuntu 22.04 can be found in INSTALL.md.

Integrating Providers

The BIQT Framework can easily be extended to integrate additional quality libraries. In general, providers must conform to the following requirements:

  • Providers must include a descriptor.json file so that metadata about the provider can be communicated to users.
  • Providers must be installed to $BIQT_HOME/providers (Linux) or %BIQT_HOME%/providers (Windows).
  • The name of the provider directory, the provider library name, and the name given in a the provider's descriptor must match.

Setting Up a New Provider

The setup_provider.py python script generates a directory structure with template files which serve as the basis for a new provider. The following example demonstrates how to generate a new provider named MyNewProvider.

Ubuntu Linux 22.04

cd ~/MyProvider
mkdir build
cd build
cmake3 -DCMAKE_BUILD_TYPE=Release ..
make
make install

Related Resources

Reference image quality libraries for face and iris images are available on GitHub.