Study Examines the Reliability of Automated Facial Recognition Software

An independent study conducted by Michigan State University (MSU) examined the reliability of automated facial recognition (AFR) software to assist law enforcement in identifying suspects in a simulation of the forensic work done after the Boston Marathon Bombings.


Media reports of the post-Boston Bombing investigation created a common perception that facial recognition technology is not always reliable because it apparently was not a factor in identifying the suspects. The published results of the MSU study challenge this perception.

In the MSU simulation, researchers used actual law-enforcement video from the bombing and searched it against a background database of 1 million law enforcement booking images. They found that the NEC NeoFace product produced a “rank one” identification—a match—of suspect number two.

“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,” said Raffie Beroukhim, vice president of NEC’s Biometrics Solutions Division.

The MSU study found that the NEC NeoFace solution consistently registered highly accurate facial matching scores in the study’s simulation.

The MSU technical paper, "A Case Study on Unconstrained Facial Recognition Using the Boston Marathon Bombings Suspects," can be accessed here.

< Prev

Item of Interest

The language barrier between English-speaking investigators and Spanish-speaking witnesses is a growing problem. (Updated 28 February 2011)