Iris is one of the most popular biometric traits. Combination of touchless scanning, stability over time and high recognition accuracy enable the use in surveillance as well as security applications.
It was shown that the iris recognition accuracy depends on the quality of the captured iris image and image preprocessing. To reduce the negative influence of illumination NIR (near infrarred) light sensing camera is recommended (Fig. 2.3). Using NIR light allows to add addition light source without influencing comfort of sensing.
Iris based identification consists of iris localization, feature extraction, and classification. One of the most successful systems achives 100 percent accuracy in controlled environments. But the localization and normalization for real life application needs to be improved. This system uses Gabor filters for feature extraction where the filtered signals are quantized over two levels. By this procedure strings of binary digits (features) are obtained. By matching the closest samples using the KNN method and a hamming distance recognition is performed.