Police Report: Face Recognition CCTV Unreliable

By Mick Meaney
RINF Alternative News 

New tests conducted by police yet again prove an increasing number of surveillance systems are unreliable.

200 people were enrolled to test CCTV face recognition software on three systems which failed to recognise 8 out of 10 people, even when the system was fed images of people standing still - one of the easiest settings often used to suggest the software is reliable.

The tests were conducted on software produced by Cognitec, Bosch and Cross Match.

The federal police in Germany who conducted the test are citing “bad lighting” as the reason for the failure and required the test subjects to be fitted with RFID tags to help with the identification process.

The final test report rightly advises against the use of the system and suggests it only be in use with the cooperation of the person being checked by the biometrics software. Currently the researchers believe that 23 false alarms per day would be an acceptable amount if only 1 or 2 real criminals were caught per week.

Such use would erode public trust in the system and is therefore being put on hold for the moment. With the right amount of light the systems only generated 0.1 per cent of false alarms by failing to recognize 4 out of 10 people, which is thought to be an acceptable rate for police work. You can view the report in German here.

Use of such systems is massively invasive as they can automatically retrieve personal information stored on a central database within seconds.

Through the use of number plate recognition, cars fitted with mandatory RFID, schools fingerprinting, biometric passports, a European biometric database and ‘Spy Drones’ monitoring street activity, we see not only Big Brother nations emerging but the creation of an entire European Big Brother continent were all movements can be tracked and recorded.

RSS TrackBack URL

Translate: