TI readies DSP-based platform for fingerprint biometrics - Embedded.com

TI readies DSP-based platform for fingerprint biometrics

Dallas — Texas Instruments Inc. on Tuesday (July 13) launched a DSP-based platform that it said will allow system OEMs to quickly evaluate different fingerprint sensors and verification algorithms.

Although fingerprint biometrics is widely recognized as a possible solution to security problems ranging from network or physical access to data security on PCs, designers must still work out implementation costs, calibrate security levels for different applications and determine acceptable rates of “false rejection” rates before a selecting a technology for a specific product.

TI calls its approach “the industry's only fingerprint sensor agnostic platform.” Ram Sathappan, biometrics solutions manager of TI's digital signal processing group, said the company's new fingerprint authentication development tool (FADT) lets system designers compare and develop a fingerprint system based on four different sensors, including Atmel's FingerChip, AuthenTec AFS8600, Fingerprint cards' FPC1010 and FPC1031.

Most current fingerprint sensors are offered either in software to run on a CPU inside an embedded system or married to one or two specific DSPs. TI's approach decouples sensor from DSP. “The fingerprint authentication market is very widespread,” said Sathappan. “It's important for customers to have an open platform.”

Composed of a fingerprint sensor daughter card and standard DSP starter kits, the tool also comes with verification software from Bioscrypt and Fingerprint cards, enabling OEMs to improve the accuracy of fingerprint authentication by cleaning up the fingerprint image before it's sent for template extraction and matching.

A flexible solution is needed not only to work with various fingerprint sensor technologies but also because fingerprints vary in a number of ways. For example, “Some people's fingers are much wetter or drier than others, depending on where they live,” said Sathappan. “One sensor won't address the global market” for most system OEMs, he added.

Since FADT is supported by many TI DSPs, designers can adapt the platform to fingerprint authentication systems and across multiple applications. Those supporting FADT include TMS320C55x and TMS320C67x DSPs.

The C55x DSP generation is used in low power, portable devices such as automobiles, gun safety locks and PDAs for an individual identification match. The C67x DSP generation, featuring a parallel floating-point architecture and efficient compiler, offers better performance best suited for one-to-many identification applications that require large databases of more than 500 prints, for example.

International Biometric Group (IBG), a New York-based independent integration and consulting firm for the biometric industry, claims there are more fingerprint solutions available than solutions for all other biometric technologies combined. The firm projects that the fingerprint authentication market will grow from $198 million in 2003 to $1.493 billion in 2008.

Still, fingerprint technology is far from the dominant biometric technology. Compared to technologies such as facial recognition and voice-scan, which can use existing acquisition devices, fingerprint biometrics' growth is “contingent on the widespread incorporation of sensors in keyboards, peripherals, access control devices and handheld devices,” according to IBG.

TI hopes new tools such as FADT can accelerate the design process while lowering the entry cost for evaluation and development of new fingerprint biometric-based products.

The new Atmel (TMDSFDCATM31), AuthenTec (TMDSFDCAFS86) and Fingerprint cards (TMDSFDCFPC31) sensor-based FADT kits are priced at $245 each, as is the previously released fingerprint card sensor FADT kit (TMDSFDCFPC10). Each kit includes evaluation fingerprint software, image capture drivers and technical documentation. DSP starter kits for C55x and C67x DSPs are $395.

For more information visit www.ti.com

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.