Computer Vision Demonstration Website

Electronics and Computer Science
University of Southampton

Iris Recognition

Irises are a good feature to use for biometric recognition as they are a protected internal organ that is easily visible. Disadvantages are that irises are very small, and are often partly occluded by the eyelids.

The demo below shows the different steps involved in iris recognition. Because the demo has an option to load you own iris images it requires you to allow it to run, by clicking yes when the security dialog pops-up.

How it works

The first stage of iris recognition is to locate an iris within an image. An Integro-differential operator is used for this, which finds all circles in an image. The sum of pixel values within each circle is calculated, and then the values of adjacent circles are compared. The iris is detected as the circle with the maximum difference from its adjacent circles. The image below shows a detected iris.

Detected Iris

The next stage is to unwrap the detected iris, which transforms the circular iris into a rectangular image. An iris code is then generated from the unwrapped iris, which can be used for comparisons.

Comparisons are done using the Hamming distance function, which compares each bit of the iris code and counts the proportion of non-matching bits, which is then normalised by the total number of valid bits. If the hamming distance between two iris codes is found to be below a given threshold then those two iris are likely to be the same.


Mark Nixon & Alberto Aguado, 2002, Feature Extraction & Image Processing, Newnes





ECS | Feature Extraction & Image Processing | © 2005 University of Southampton

University of Southampton