Officially, no one’s saying anything, but a West Virginia University professor says there are many tools in biometrics technology that may have played a role in discovering the identity of the masked Islamic State fighter who appears in videos of the violent killings of foreign hostages.
Last week, United States and British news outlets identified “Jihadi John” as Mohammed Emwazi, a Kuwaiti-born man who was raised in the United Kingdom.
At a briefing in September 2014, FBI Director James B. Comey said that the U.S. had determined the identity of the masked militant, but would not disclose his name nor the tools and intelligence used. Due to the ongoing investigation, U.S. officials did not comment on last week’s revelation.
Details about how U.S. intelligence agencies made the identification are not publicly known, but Thirimachos Bourlai, assistant professor in the Benjamin M. Statler College of Engineering and Mineral Resources, says there are many ways to narrow down a suspect pool using biometrics, which is the science of identification using unique biological characteristics such as fingerprints, iris patterns and facial recognition.
“Typically, once biometric data is gathered, law enforcement and intelligence agencies can authenticate an individual’s identity or figure out who someone is by scouring a database for a match,” Bourlai said.
Bourlai explains that one of the search engines that can be used is called “blind search,” in which each query image, such as face, iris or fingerprint, is compared with all the gallery images of enrolled individuals without utilizing other intelligence such as demographic information.
However, to help narrow down the field of suspects, a more efficient search engine called “filtered search” can be used. In this type of search each probe is compared with gallery images of a more focused pool of individuals such as “only male” or “only male with a beard.”
He also says that identifying the man known as “Jihadi John” probably presented unique challenges in the absence of fingerprint data or high-quality full-frontal face images.
“In the videos, the man’s face is almost completely obscured with only the periocular region – the eye and eyebrow area – visible,” Bourlai said. “Beyond that, the video was taken from a distance and the face images – or frames – available were off-angle and of low image quality.”
Bourlai’s research in facial recognition focuses on how image quality – such as low illumination, watermarks or other image artifacts – affects the performance of facial recognition systems. He is also working on how to better match visible face images against face images captured either with the same or different cameras, which are within or outside the visible spectrum, such as near infrared, short-wave infrared, mid-wave infrared (thermal) and long-wave infrared (thermal) cameras.
The FBI’s Criminal Justice Information Services Division, located in Clarksburg, is the home to the Biometric Center of Excellence. CJIS and the BCOE collaborate with a variety of agencies and partners. In academia, the BCOE has established a cooperative agreement with WVU to engage in applied research and facilitate engagement with other academic institutions.
Bourlai is available to offer commentary to the media. He can be reached at 304.293.4326 or Thirimachos.Bourlai@mail.wvu.edu.
-WVU-
ms/03/03/15
CONTACT: Mary C. Dillon; WVU Statler College of Engineering and Mineral Resources 304.293.4086; Mary.Dillon@mail.wvu.edu
Follow @WVUToday on Twitter.