Automated recognition of people by the pattern of the iris in the eye offers major advantages over conventional identification techniques.
The random and highly complex patterns in the structure of irises are even more individual than fingerprints and these features lend themselves more readily to rapid checking against a computer database. Iris recognition systems require very little cooperation from the subject and are virtually impossible to deceive.
Dr. John Daugman at Cambridge University has developed computer algorithms that enable the unique features of an iris to be encoded in as little as 256 bytes, which already enables the identity of an individual to be established, verified or denied by comparison with reference iris patterns, at a rate of 40,000 persons a second using ordinary computer equipment. Simple dedicated hardware can speed up this search to 160 million people a second, believes Dr. Daugman.
One of the first commercial applications of Dr. Daugman's technology is to identify users of ATMs (automatic teller or cash dispensing machines) through the use of a built-in "smart" camera now being developed by Sensar of the USA for Oki, Japan's largest manufacturer of ATMs.
Manufacturing and user licences are also being negotiated in Germany, France and the UK, where British Telecom and others are studying the concept.
The complex mathematics needed to encode and compress the large quantity of digital data contained in the image of an iris into a small data-storage volume - while enabling the presented image to be compared to a reference image with extremely low error rates (said to be many hundreds of times less than other biometric identification Systems) - is at the heart of the technology.
Our picture shows Dr. Daugman studying an iris image on the lower monitor screen while the upper screen displays some of the mathematics involved in the image encoding process.
The structure of an iris remains stable over decades. The system also detects the small variations such as constriction and dilation of the pupil, which are characteristic of a living eye, so that it cannot be deceived by, for instance, presenting a fake eye or photograph.
In addition to ATMs, the wide variety of potential uses for the technology includes control of access to buildings, secure areas and so on, and verification of personal documents. Licences are also currently being negotiated for car anti-theft devices, and for use of the iris pattern as an encryption key for securing telecommunications over non-secure channels.
For more information contact:
Dr. John G Daugman
University of Cambridge Computer Laboratory
Pembroke Street, Cambridge, England, United Kingdom, CB2 3QG
Tel: +44 1223 334501
Fax: +44 1223 334679