AuthenTec Speeds Fingerprint Matching
Security breaches and transaction fraud have catapulted the need for highly secure identification and personal verification firmly into the limelight. Biometrics, an automated method of recognizing a person based on physiological or behavioral characteristics, is quickly emerging as highly preferable to identification systems based on passwords, PINs, and swipe cards.
Analog Devices processors have been used for countless applications that involve image processing, the underlying technique used for virtually all biometrics systems. Moreover, the company's Blackfin� embedded processor brings industry-leading microcontroller features that made it the best choice for AuthenTec's FingerLoc� AFS 8600 embedded fingerprint matching reference design. AuthenTec, a fingerprint-identification leader, needed a low-cost embedded processor solution that did not sacrifice identification-time performance.
AuthenTec wanted a technology partner who could help them "take biometrics into areas where we have never been able to go before," while meeting the company's key needs for low cost, fast identification time, and good power management. Powered by a Blackfin processor, fingerprint identification algorithms and dataflow that previously required more than a full second to execute are executed in less than less than one-half second.
Virtually all biometric techniques are implemented using a sensor to acquire raw biometric data from an individual, feature extraction to processes the acquired data to develop a feature set that represents the biometric trait, pattern matching to compare the extracted feature set against stored templates residing in a database, and decision-making whereby a user's claimed identity is authenticated or rejected.
AuthenTec's Blackfin processor-driven sensor technology uses a ridge-flow pattern-matching approach in a realtime closed-loop control system to provide exceptional image quality and repeatability to reduce the cost of commercial fingerprint identification systems. As an "embedded core that can do good signal processing," the Blackfin processor, according to AuthenTec, was a good architectural fit for the application.
AuthenTec uses the Blackfin processor, an ADSP-BF531, to pre-process an acquired fingerprint, introducing frequency, amplitude and phase modifications to deliver the best possible image to the pattern-matching algorithm. AuthenTec's TruePrint� sensor technology is more resistant to distortions such as cuts or scratches and other day-to-day variations of the finger. Its pattern detection technologies capture useful fingerprint images by reading the patterns from just beneath the skin surface. This is in contrast with all other fingerprint sensors, which read the patterns at the actual skin surface. The result is an industry-leading 0.1% EER (Equal Error Rate) at a medium security level.
The AuthenTec system also uses feedback to improve enrollment over time - each time a user fingerprint is sampled by an AuthenTec sensor, that user's template database is augmented and improved.
While Blackfin's processing muscle easily provides more than enough signal processing MIPS to execute the AuthenTec algorithm, it was the processor's equally impressive microcontroller features that made it the ideal engine for the AuthenTec application. The company was especially attracted to the processor's power management features, including the dynamic clock rate shifting capability.
Fingerprint identification is a natural choice for security in a variety of portable devices such as PDAs, laptop computers, and cellphone markets, but not at the cost of battery performance. At 1/3 to half the power of competing embedded processors (The Blackfin processor's Enhanced Dynamic Power Management brings power requirements as low as 0.15mW/MMAC at 0.8V), the AuthenTec solution is highly attractive for these kinds of power-sensitive applications.
The Blackfin processor's tiny footprint (12 mil by 12 mil package) is an added bonus for size-constrained fingerprint-based security devices.
High-performance at low cost
Because price point is a crucial criterion for reference designs, bill-of-materials (BOM) costs are scrutinized carefully, and at just $5, the Blackfin processor makes up a far smaller percentage of a BOM than does the typical embedded processor. The highly integrated Blackfin processor enabled AuthenTec to drop its reference design BOM costs to less than one-half of previous solutions, and enabled the company to use more cost-effective memory configurations. In fact, Authentec identified the low cost of the Blackfin processor as a "dramatic and dominant" reason for choosing this embedded engine.
Analog Devices' Blackfin processor, with its industry-leading microcontroller features, was clearly the best choice for fingerprint-identification leader AuthenTec. The Blackfin processor enabled AuthenTec to achieve real-time fingerprint identification in less than one-half second, a previously unattainable feat. With The Blackfin processor, AuthenTec enables its OEM customers to quickly "put a finger" on highly secure identification and verification systems-an absolute must in today's security conscious world.
For more information, visit AuthenTec.
Read more Customer Case Studies like this...