Biometric CMOS sensor aids fingerprint reconstruction

In its annual special report, The Gartner Group (Stamford, CT) reported on the ten technologies to watch in 1999. According to the report, one of the technologies is biometrics--the ability to identify an individual by recognizing one or more unique physiological characteristics, such as a fingerprint, an eye`s iris, or a facial feature.

Biometric CMOS sensor aids fingerprint reconstruction

In its annual special report, The Gartner Group (Stamford, CT) reported on the ten technologies to watch in 1999. According to the report, one of the technologies is biometrics--the ability to identify an individual by recognizing one or more unique physiological characteristics, such as a fingerprint, an eye`s iris, or a facial feature.

Whereas traditional electronic fingerprinting systems use optical systems to capture fingerprint images and offer a high degree of image quality, image distortion can still be introduced by the optical components. In addition, optical systems are susceptible to physical shock and impact. Although optical systems are becoming smaller, they can never be as small as a silicon device due to the optics and focal-length requirements. Although many companies have introduced CCD- or CMOS-sensor- based systems to capture fingerprint images, bright sunlight can saturate such sensors, making it extremely difficult to use these systems in outdoor environments.

To overcome this problem, Thomson-CSF (Totowa, NJ) has developed a thermal CMOS sensor, the FC15A140 FingerChip, that incorporates special layers within the chip to reduce ambient light problems. The first production version of the 1.5 x 14-mm CMOS FingerChip sensor comes in a ceramic package, can capture several successive images, and can reconstruct them into a full fingerprint via a Thomson proprietary algorithm. With this scanning technique, the sensor is capable of imaging a fingerprint ten to 15 times larger than its surface.

To incorporate this sensor into biometric systems, Thomson-CSF has initiated a FingerChip partners program that has resulted in several OEM PC-based modules and fingerprint-recognition products that couple the device to DSP- or RISC-based products. For example, GEZ Microsystems (Oakville, Ont., Canada) has released a family of stand-alone fingerprint-enrollment and verification modules using the FingerChip and the Motorola (Tempe, AZ) DSP56309 processor. By incorporating the sensor with a processor, GEZ Microsystems aims to address home security, corporate access, and automobile security applications. Occupying 1.5 x 1.7 x 0.5 in., the sensor modules use a proprietary verification algorithm and fingerprint templates to recognize scanned images. It can be attached to a PC via an RS-232 interface.

For PC-based development over the ISA and PC/104 buses, Hopkins Imaging (Duarte, CA) is offering both ISA and PC/104 interfaces. Using a ribbon cable between the bus card and a FingerChip board allows the sensor to be mounted several meters away from the host CPU. For systems developers, Hopkins also supplies a 32-bit dynamic link library and a Microsoft Windows demonstration program. For faster acquisition, the company is currently developing a PCI bus interface that will be based on a custom interface chip.

Like Hopkins Imaging, Oxford Micro Devices (OMD; Monroe, CT) also offers a PC-based interface for the Thomson-CSF device. To digitize the images from the FingerChip, OMD has developed a small module that interfaces to the company`s A236 PCI-based evaluation board over a serial link. To digitize analog signals from the FingerChip, the module incorporates an Analog Devices (Norwood, MA) 9280 analog-to-digital converter. After digitization, a custom-developed programmable array logic device converts the parallel digital information to a serial data stream that can be interfaced to the A236 evaluation board. Here, the serial data stream is reconverted into a parallel data stream for processing by the company`s A236 parallel-video DSP device.

Using the evaluation board, images can be captured, reconstructed, and displayed on a video monitor in real time using OMD software. According to Steven Morton, OMD president, the company is looking for large-volume embedded applications that use the FC15A140 FingerChip and the A236 processor. To foster development, the company supplies a fingerprint-development kit with board-level schematics and processor development tools that include assembler, linker, loader, and debugger software.

Such embedded applications are likely to become popular as sensor and processor prices continue to fall. Already, Cezam (Montigny les Cormeilles, France) has embedded the Thomson-CSF FingerChip, host CPU, and memory to store 300 fingerprints into its FingerChip mouse, a USB-compatible mouse that is supplied to developers with a PC-hosted software development kit.

More in Non-Factory