Can Iris Scanning Technology Tell the Difference Between Twins?
If security is going to become completely dependant on biometric identity technology, the technology must be as close to 100% accurate for all people. As of now, there are still too many things that could go wrong. What if, when preparing an iris image database, for example, the initial scan was flawed? Or what if someone was able to hack into the database and copy the iris bitmap, confusing you with someone else or stealing your identity altogether? Something that may throw the system off completely is the case where two or more individuals have the same genetics and physically look identical. Will these biometric systems be able to distinguish one person from the other in a pair or twins?
In a recent post at ScienceBlog.com, the author writes about new studies going on at the University of Notre Dame’s Department of Computer Science and Engineering. Each year, the town of Twinsburg, Ohio (go figure…), hosts the largest gathering of twins for its annual convention that features food, entertainment, a twins parade, and a golf tournament. Kevin Bowyer and Patrick Flynn, two researchers at Notre Dame who have been working with biometrics since 2001 have been attending to study the pairs and record biometric measurements like face photos, thermograms, 3D facial imaging, iris scans, and documentary video.
Working from two grants from the Federal Bureau of Investigation, the two researchers assemble a team to walk the convention and record biometric samples of twins’ faces and irises. Volunteers are seated at the center of a half-circle surrounded by five cameras that shoot high-resolution color photographs from various angles. They pose for iris and 3D face imaging cameras, and the collected data is eventually used in research studying twins and biometric recognition.
Bowyer and Flynn show paired images of irises to a group of volunteers at the university and record their opinion as to whether the irises in the pictures belong to twins or not As it turns out, their volunteers can classify pairs of twins with an 80% accuracy, far more than you might initially expect.
But even twins have differing irises. The team plans to analyze more data to look at the potential of creating new types of automated iris image capture and analysis.
Twins are intriguing research subjects for Notre Dame biometircs researchers
Related articles:
- Eye Tracking: Iris Scanning Is Coming, Like It Or Not
- Iris Scanning Goes Big
- Eye Tracking: Iris Scanning to be the Ultimate Reference Point?
- Passengers Accept Iris Recognition Technology in Major Airports
- Beware of Problems With Iris Recognition
- Eye Tracking: Uncertainty in Iris Localization
- The All-In-One AOptix Face and Iris Recognition System
- Iris Recognition Becoming Critical Part of NYC Police Database
- Iris Recognition is Advancing. Could It Help Eye Tracking?
- Spreading Iris Tracking Technology to Every State