Learrnn-ing curve for BAE Systems

BAE Systems said the first results of its technology partnership with electronic design specialist Celoxica will emerge next month in the biometrics and image-processing sectors.

The UK defence giant and Celoxica will demonstrate new automatic fingerprint-matching technology and a high-speed, real-time video-processing system in the US.

Oxfordshire-based Celoxica and BAE signed a partnership deal in February to develop advanced processing capabilities for electronic systems.

The two companies plan to tap into the market for hugely increased processing power able to make real-time comparisons in areas such as biometric identification, transaction fraud detection and machine vision.

Biometric identification systems – which use a unique characteristic of the human body to verify the identity of individuals – are currently one of the hottest areas of technological research thanks to security fears.

However, the effectiveness of biometric systems has been questioned because of the massive demands for accuracy and speed involved in gathering and processing data from millions of people.

The BAE/Celoxica work is based on neural network technology developed by BAE called Learrnn (Logic-enabled asynchronous rapid robust neural network).

According to BAE, Learrnn is highly effective at real-time filtering and interpretation of 2D and 3D digital data. It can operate at high speeds to perform a variety of pattern recognition, fault detection and diagnostic applications, the company claimed.

Celoxica is working with BAE to help design Learrnn into specific products. The technology is compact and cost-effective to imple- ment compared to other similar systems, the pair claimed.The first fruits of their co-operation will be demonstrated at a conference in California.

The fingerprint-matching engine can match a print against a database of up to 800 in less than 50 milliseconds, said BAE and Celoxica. The technology platform has been designed to be able to support very large fingerprint databases, they added.

The second product is a real-time video filter able to deal with very high frame rates. BAE and Celoxica plan to market the applications in sectors such as security surveillance, mobile telecoms, broadcasting and the emerging area of video-based automotive safety systems.

BAE said it had seen a major growth in demand for technology in the imaging and security sectors. Other key markets for the Learrnn-based systems could include medical and industrial automation applications.