printer friendly version

Facial biometrics to fight crime

Facial biometrics can be deployed in the fight against the increasing number of armed robberies in South Africa.

The number of armed robberies has escalated in the last six months to levels where it is becoming increasingly difficult for authorities to effectively analyse the volumes of video footage produced by CCTV camera systems before and after robberies.

In previous years, armed robberies in public areas was something normally associated within the banking environment. Today shopping malls are a prime target as well as industry. In Port Elizabeth, industries dealing in precious metals for automotive catalytic converters have become targets of armed robberies as well as outlets selling bulk cigarettes and liquor. Clearly the pattern of armed robberies is across the board and the subsequent additional amount of video footage generated is more than what can easily be handled.

In reviewing such footage, the objective is to identify the criminals’ active in these armed robberies as well as linking them with what may be called the passive participants who are also involved with these robberies. The passive participants normally do not hide or disguise their identity and can be vital links in establishing gang relationships.

The manual process of reviewing CCTV video footage, reliant on the human factor to memorise and tediously refer to other incidents and recorded material in an endeavour to identify and establish links between these persons is a very time consuming process.

The process becomes a more tedious task when individual incidents need to be referenced back to more than 1000 known perpetrators on record and the association of passive participants becomes a near impossible task.

Biometric facial recognition technology exists which can assist in the evaluation of passive and active participants in criminal activities. In using facial recognition software, investigators can:

* Develop and maintain an electronic database of known criminals and participants involved in robberies.

* Compare and identify facial images from new and existing CCTV footage of bank or similar robberies against this database in a matter of seconds with best fit and similar comparisons while indicating the level of probability.

* Further classify images in the database.

* Simply enrol new images into the database to further complement the scope of the comparisons.

* Control a database of millions of images which can be distributed over a wider area network, WAN, to enable processing and evaluation by numerous client work stations around the country.

The facial recognition software can interrogate previously recorded video footage in live mode and identify facial images against those stored in a database. This has the advantage whereby new footage can be evaluated electronically quickly to identify known criminals and passive participants.

This software can be deployed in public places to evaluate live footage from installed CCTV cameras as an early warning to banks and malls of the presence of listed people on the data base. Facial biometrics thus enables a new generation of computer-assisted video surveillance.

For more information on Cognitec’s FaceVACS facial recognition applications please contact Modular Communications.

For more information contact Cliff Rose, Modular Communications, modular@worldonline.co.za

Share this article:

Further reading: