The first year that we looked at biometrics, we had some of the most fascinating devices it ever has been my pleasure to test. We had facial recognition, fingerprint scanners and palm geometry devices. There were flashing lights from the facial recognition devices. Another year we had one of the first vascular devices available. This device looked at the veins just under the skin. There also have been retinal scanners. Overall, I think, we've seen it all.
Last year, something started to happen. Almost all we saw were fingerprint scanners - ranging from very bad to very good biometric devices. As I wrote not long ago, the devil is in the software. Good biometric software paired with a competent scanner can make a very solid fingerprint scanner. However, there has been an ongoing debate about which kind of biometrics is the best. For a while it looked as if fingerprints were going to carry the day. However, facial recognition started to gain ground - with research being driven by the convenience of being able to scan without the subject realizing it and the need to identify terrorist suspects at airports and other travel spots.
However, fingerprint scanners were starting to boast 99.9 percent accuracy while facial recognition was pegged at just under 90 percent. Also, two-dimensional facial recognition was plagued by lighting challenges and that affected accuracy significantly.
Facial recognition research goes on apace, testing the use of multiple cameras, different lighting requirements, two- and three-dimensional rendering, etc. Meanwhile, fingerprint scanners are starting to become very affordable, as well as accurate. Functioning quietly in the background are the palm geometry devices that are used mostly for physical access. Additionally, some developers of biometric software are starting to look at mixing different types of devices, so you might have facial recognition paired with fingerprint scanning.
So, exactly what should you look for when you buy biometrics? First, don't let budget constrain your decision to use biometrics or not. If you select the right application and the appropriate type of device, you can put together a very respectable authentication tool at a very reasonable price. The key is a three-step process. First, why do you want to use biometrics in the first place? Is what you are protecting worth the cost and the trouble? In other words, can the biometric device that opens the data center door authenticate to the user's computer too?
Second, once you've decided that biometrics is the answer, what are you going to do with it and what is the appropriate technology? With today's fingerprint-scanning software, that may be the answer for most logical access applications. If you need to open doors, though, you might want something a little more robust. Consider that there is a lot more behind that data center door than there is under the hood of your laptop.
Finally, once you've decided that you need biometrics and you've picked out the appropriate technology, decide on your budget. You can control budget by deciding what your mix of risk, technology and convenience needs to be. That will, perhaps, narrow the application field. Not all users need biometrics. There are other ways to get strong authentication. Save the biometrics for the heavy applications. It's not just the cost of the devices. It is the total lifecycle cost with which you must contend.
This month, our testing was about as straightforward as it gets. We applied the products in the same way that I've described your selection process. Then we put the products through their paces. The results follow.
Michael Stephenson contributed to these reviews.
Biometric authentication: Growing in sophistication and useIn the United States, biometric authentication technology has taken huge leaps as a result of the September 11 attacks. According to the International Biometric Group (IBG), in 2009, fingerprinting methods made up about 45.9 percent of all biometric methods used in government or industry. Facial recognition was second at 18.9 percent. This technology is used less frequently due to the inaccuracy rate.
However facial recognition technology is improving and is becoming the choice for government and industry due to its ability to authenticate someone without the target being aware that authentication is taking place. Professor Mohamed Abdel-Mottaleb of the University of Miami and his colleagues have designed a new system that is able to photograph a person's face and compare it with previously stored images of the same person with 95 to 100 percent accuracy. At the 2009 IEEE International Conference on Image Processing in Cairo, Abdel-Mottaleb presented findings on his team's research. He stated that his work is, "satisfying, especially when you know that what you're doing has real-world applications that will benefit people and enhance personal security."
Abdel-Mottaleb and the researcher's system is capable of using 3-D or 2-D images of a face. The former method focuses on 3-D images of the unique structure of the ear for authentication. In a lab setting, the 3-D image technology has a 95 percent success rate. However, the process takes more computational power and time compared to 2-D image processing.
To combat the resources needed for the 3-D method, Abdel-Mottaleb has focused his 3-D system to ignore distinguishable features while highlighting other key areas of the face. The 2-D imaging method, in the lab setting, has a success rate of 100 percent. However, the 2-D method is sensitive to changes in lighting.
Biometric authentication is being used by law enforcement, the military (abroad and stateside), government, as well as in industry. The research conducted by Abdel-Motteleb has improved facial authentication technology. However, he and his team are constantly looking at ways to improve on their existing research.
Vincent Lally is a student at Norwich University.