Partial iris and recognition as a viable biometric scheme

摘要

Iris recognition is a well-studied biometric technique that is both reliable and convenient. The paper focuses on the hypothesis that due to the rich structure of the iris data, a portion of the iris data could render results comparable to those achieved by processing the full data. Experiments are conducted on partial iris data using Daugman's algorithm. Using the phase orientation of Daugman's algorithm, only half of the iris data was examined in a series of tests. Identical irises were identified using partial data in most cases. For those cases that a match could not be found, the similarity score for identical irises was still significantly higher than that for different irises. The results support the promise of making the iris recognition a covert technique with many applications in security.