Biometric entropy of iris and retina

eye imaging, entropy, iris,  retina, biometrics, human recognition, security

Paper will be describing determination of the amount of entropy in biometric iris
and retina for biometric purposes. In other words, it is a combination of all
possible occurrences minutia. Based on other experimental findings, formula is
further modified and made it more precise.

Biometric identification systems have become a usual part of requirements for
increasing security needs, also in strictly protected areas such as nuclear
plants, military facilities, and scientific laboratories; however they are used
in common life as well. Iris or retina recognition is not a new idea but if we
consider this as a multimodal system this gives us a totally different view to
the security and reliability.

Biometric entropy describes the inherent variability in biometric samples in the
population. It can also be understood as the information content of biometric
samples is related to many questions in biometric technology. For example, one of
the most common biometric questions is that of uniqueness. Such a measure is
important for the performance of biometric system, as a measure of the strength
of biometric cryptosystems and for privacy measures. It also is relevant for
applications such as biometric fusion, where one would like to quantify the
biometric information in each system individually, and the potential gain from
fusing the systems.

In our case the computation of the iris and retina biometric entropy is performed
in several ways, similar to fingerprints. We studied every pixel of the retinal
image and checked whether there can be a split point of the vessels (bifurcation)
or their termination (end point). Theoretically, we could say that it can occur
anywhere in each pixel. As we know from images and physics of the retinal blood
vessels, it is not possible practically. So, if we assume that in one pixel
a bifurcation occurs, the next bifurcation can't be located in its close
surroundings. Using this assumption, we can predict total number of minutiae in
one particular retinal image. For each occurrence, we remember its properties
(position in the image, the angle of the blood vessels, the minutiae type and the
thickness of the blood vessels).

Practical calculation of the entropy from properties of the eye is similar to the
theoretical way. It is needed to find out the average occurrence of features in
the retina. That means bifurcations, vessels crossing and termination. It will
also depend on the image resolution and the quality of imaging systems.