On the Role of Interpolation in the Normalization of Non-ideal Visible Wavelength Iris Images

The growth in practical applications for iris biometrics has been accompanied by relevant developments in the underlying algorithms and techniques. Along with the research focused on near-infrared (NIR) cooperatively captured images, efforts are being made to minimize the trade-off between the quality of the captured data and the recognition accuracy on less constrained environments, where images are obtained at the visible wavelength, at increased distances, over simplified protocols and adverse lightning. This paper addresses the effect of the interpolation method, used in the iris normalization stage, in the overall recognition error rates. This effect is stressed for systems operating under less constrained image acquisition setups and protocols, due to higher variations in the amounts of captured data. Our experiments led us to conclude that the utility of the image interpolating methods is directly corresponding to the levels of noise that images contain.