The Department of Homeland Security's latest surveillance tool is the Biometric Optical Surveillance System, or BOSS, and consists of two cameras that are capable of taking stereoscopic images of people's faces. Facial recognition technology matches the 3-D pictures captured and compares them to images stored in the system's database like driver's license photos.
Privacy advocates have expressed their concerns regarding the technology, and RT's Adriana Usero explains the latest developments in surveillance. The Department of Homeland Security's latest surveillance tool is the Biometric Optical Surveillance System, or BOSS, and consists of two cameras that are capable of taking stereoscopic images of people's faces. Facial recognition technology matches the 3-D pictures captured and compares them to images stored in the system's database, like driver's license photos. Privacy advocates have expressed their concerns regarding the technology, and RT's Adriana Usero explains the latest developments in surveillance.
U.S. testing crowd-scanning facial recognition system
The U.S. Department of Homeland Security will test its crowd-scanning facial recognition system, known as the Biometric Optical Surveillance System, or BOSS, at a junior hockey game this weekend, according to the Russian news agency RT.
With assistance from the Pacific Northwest National Laboratory, DHS will test its system at a Western Hockey League game in Washington state. The test will determine whether the system can distinguish the faces of 20 volunteers out of a crowd of nearly 6,000 hockey fans, to evaluate how successfully BOSS can locate a person of interest.
According to DHS, BOSS technology consists of two cameras capable of taking stereoscopic images of a face and a back end remote matching system. Stereoscopic images are two images of the same object, taken at slightly different angles that create an illusion of three-dimensional depth from two-dimensional images.
The cameras transfer the pair of images to the remote matching system by way of fiber optic or wireless technology. The system then processes and stores the two images into a 3-D signature, which is the mathematical representation of the stereo-pair images that the system uses for matching.
Using the BOSS facial recognition algorithms, the signature is matched against a locally stored database created from volunteers, using a combination of mathematical and statistical analysis.
BOSS is capable of capturing images of an individual at 50-100 meters in distance. The system can capture images of subjects participating from a specific distance, or be set up in a way that tracks and passively captures frontal face images of an individual as he/she moves in front of the camera.
As reported previously in BiometricUpdate.com, a $5.2 million contract for BOSS was awarded to Electronic Warfare Associates, a U.S. military contractor.
Recently the system was not deemed ready since it could not achieve 80 to 90 percent identification accuracy at a distance of 100 meters and could not process and identify images in less than 30 seconds against a biometric database.
This weekend’s test will attempt to rectify this deficiency. If the test succeeds, the technology conceivably could be used at international crossings and other ports across the United States patrolled by the department.
The U.S. Department of Homeland Security will test its crowd-scanning facial recognition system, known as the Biometric Optical Surveillance System, or BOSS, at a junior hockey game this weekend, according to the Russian news agency RT.
No comments:
Post a Comment