Personal Identification System using Hand Geometry and Iris Pattern Fusion

This research proposes a fusion of two biometric systems using hand geometry and iris pattern for personal identification. Each individual system uses simple features which make calculations easy and fast. Four normalization methods at the matching score level are tested in this research. They are min-max method, z-score method, possible min-max method and possible min-max with minimum score selected method. The CER of each unimodal system are 8.94% for the hand geometry system and 8.45% for the iris system. The best CER, 1.67%, of hand geometry and iris fusion is from the possible min-max with minimum score selected normalization method