The FBI will be pouring USD $1 billion into a new face recognition system that is capable of identifying someone under various conditions with as much as 92 percent accuracy.
The Next Generation Identification (NGI) pilot program was launched in 2010, and according to the results, the best algorithms were capable of narrowing down someone’s identity 92 percent of the time from a pool of 1.6 million mugshots.
Even if a person is not directly looking at the camera, the technology is still able to match the person to the available mugshot in the database using various biometric analyses. The algorithms can analyze features of front and side views of mugshots, create a 3D model of the face, and rotate the model to as much as 70 degrees to match the angle of the face in a photo.
However, the system also has its limitation, because identifying faces from pictures in low light environments have proven to be an obstacle.
Although the technology will allow law enforcements to identify and arrest criminals faster, privacy advocates are worried that innocent people might end up in the FBI’s surveillance database if they are caught in the photos of suspects.
The FBI has already partnered with issuers of state drivers’ licenses to obtain posed images for photo comparisons in their NGI program. The program is expected to be in full force come 2014.
The Next Generation Identification (NGI) pilot program was launched in 2010, and according to the results, the best algorithms were capable of narrowing down someone’s identity 92 percent of the time from a pool of 1.6 million mugshots.
Even if a person is not directly looking at the camera, the technology is still able to match the person to the available mugshot in the database using various biometric analyses. The algorithms can analyze features of front and side views of mugshots, create a 3D model of the face, and rotate the model to as much as 70 degrees to match the angle of the face in a photo.
However, the system also has its limitation, because identifying faces from pictures in low light environments have proven to be an obstacle.
Although the technology will allow law enforcements to identify and arrest criminals faster, privacy advocates are worried that innocent people might end up in the FBI’s surveillance database if they are caught in the photos of suspects.
The FBI has already partnered with issuers of state drivers’ licenses to obtain posed images for photo comparisons in their NGI program. The program is expected to be in full force come 2014.