Standardized Computer-On-Modules (COMs) offer users up-to-date, application-ready processor technologies and have thus firmly established themselves in the embedded world. On the one hand, Intel ® has occupied the low power area of x86 with its Intel ® Atom™ processor family and on the other hand, addresses the high-end market for power hungry applications with its 3rd Generation Intel ® Core™ processor family with quad-core and dual-core. Today, extremely powerful modules are implemented in the established COM Express™ form factor. However, further improved graphics, video and computing performance is not everything. With the revision from Type 2 pin-out to Type 6 pin-out in the latest COM Express™ specification, up to three Digital Display Interfaces (DDI) and USB 3.0 are now supported. The power dissipation of such modules is, however, simply too high for passively cooled systems.
The Qseven™ standard, which defines modules in a compact form factor of 70 mm x 70 mm, bridges the gap from x86- to ARM ®-based modules. In addition to support for various x86 processors, with the Qseven™ Specification Revision 1.2 support was added for power-efficient ARM ®/RISC architectures. Thus, the basis for a series of particularly compact embedded modules with a number of different types of processors for a wide range of applications has been established. Additional fields of application are opened up through the integration of a digital signal processor (DSP). DSPs can manage the demanding data and image analysis tasks in real-time and thus significantly reduce the load on the integrated ARM ® processor. The use of DSPs is especially worthwhile in applications such as imaging, demanding visualization and media in, for example, industrial automation, medical and measurement technologies as well as in the fields of transportation and safety engineering.
One of the most promising applications for Computer-On-Module (COM) with high computing power and demanding digital signal processing are biometric identification methods, which have experienced an enormous boom in the last few years. The rapid measurement of biological characteristics and their analysis with reasonable effort in high quality are only possible as a result of technological progress in image processing and analysis technologies. The basis of biometric methods for computer-assisted identification of human beings is based on biological features such as fingerprint, handprint, hand/finger geometry, face, eye (iris and retina) as well as behavior-specific features such as voice, typical body movements, signature or the rhythm of keyboard strokes.
A biometric identification system integrates a sensor component, for example a video camera, which captures a biometric sample. By means of complex algorithms, any superfluous information provided by the sensor which does not contribute to biometric identification is filtered out. After establishing a biometric reference template, where biometric features have been stored for the purpose of a comparison, the submitted biometric sample is compared with the template. The system now determines whether the score, which designates the degree of similarity between the submitted biometric sample and the previously stored reference template, is sufficiently high for verification of identity.