§ Iris Recognition Application
The iris recognition application is free, open-source, cross-platform and developed in C++ and Qt.
What is its use?
The application provides a simple and intuitive interface for reading in images of eyes, comparing them, and verifying their identity.
A simple screenshot
Below you can see the application after it has read in an image of the eye, and detected the location of the pupil, iris and eyelids:
What kind of features does it have?
Some of the notable features are:
- Reads in an 8-bit greyscale image of the eye
- Auto-detects the pupil, iris and eyelids to a high degree of accuracy
- Allows you to manually specify the location of the iris and eyelids
- Can store iris bitcodes in a text database and compare them to others
In general, the application implements the algorithms and methodology invented by Professor John Daugman (patent, 1, 2). The auto-detection parts use various other methods.
§ Download
The iris recognition application is available under the GPL licence for non-commercial research purposes on Linux, Mac and Windows desktops.
Source tarball
Mercurial repository
To get the mercurial branch simply run:
hg clone http://projectiris.co.uk/iris
Alternatively, you may view the changelog or browse the source.
§ Installation
In short (for Linux):
tar zxf iris-version.tar.gz
cd iris-version/
qmake
make
./iris
You can sample the application with two of the images of eyes available in that directory.
§ Report
A paper was also completed on the project which is available here:
§ Authors
This application is the work of a few MSc Computing Science students at Imperial College London who created the application as a result of their group project. They are:
- Michael Boyd
- Dragos Carmaciu
- Francis Giannaros
- Tom Payne
- William Snell
They worked under the supervision of Professor Duncan Gillies.