Personenidentifikation mittels Principal Component Analysis von Irisbildern

Biometric identification based on iris measurements is a successful approach for biometric identification, which is, at the time, practically only accompished by means of feature extraction via Gabor-Wavelets (and in this form covered by existing patents). Dominating the market are systems using John Daugman’s Iris Code2. Even though these systems combine high recognition rates with high recognition speed, a system not using patented technologies whould be interesting and desireable as a freely available alternative. This work examines inhowfar the Principal Component Analysis (PCA) is suitable as an alternative to feature extraction via Gabor Wavelets. Throughout this work the design of a biometric system, divided into its most important components - like image acqusition, image emhancement, segmentation, localisation and finally the Principal Component Analysis - is decribed and the performance evaluated.