NeoFace identifies Dzhokhar Tsarnaev in MSU marathon bombing simulation study
NEC’s NeoFace facial recognition software was recently able to identify Dzhokhar Tsarnaev, a suspect from the recent Boston Marathon Bombing, from footage of the event as a part of a study conducted by Michigan State University that simulated post-bombing forensic work.
The focus of the study was to examine the reliability of automated facial recognition software to assist law enforcement in identifying suspects in a simulation of an event like the recent marathon bombing.
Following the initial identification of the Tsarnaev brothers – the two accused of the bombing – there were many reports to suggest that facial recognition software was unhelpful and did not play a role in identifying the suspects. According to NEC, the results of this study challenge that assertion
As soon as the story of the marathon bombings broke, BiometricUpdate published a feature which explored the use of facial recognition to identify the Boston bombers and discussed the technology’s benefits and limitations in identifying a suspect.
In the MSU simulation, researchers used actual law enforcement video footage from the bombing and searched it against a background database of 1 million law enforcement booking images. According to NEC, the company’s NeoFace software produced a “rank one” identification – a match of suspect number two, Dzhokhar Tsarnaev.
“As conditions in the study were simulated to be representative of actual crime scene situations, including the limited, poor quality images available to the investigators of the Boston Marathon Bombings, the strong performance of the NEC solution is significant,” Raffie Beroukhim, vice president of NEC’s Biometrics Solutions Division said.
According to a recent CNN/Time/Orc poll, 79% of Americans are in favor of using facial recognition at various locations and public events, and 81% support expanded camera surveillance on streets and in public places.
The MSU technical paper on evaluating automatic facial recognition technology can be found on the MSU website.