One female sought in connection with the March 13 homicide of Deborah Burton, a 62-year-old from Maplewood, N.J., is shown on a snow-covered street of Newark, N.J., shortly after the killing. A new paper published in the Journal of Forensic and Legal Medicine proposes a new way to estimate biometric information based on a comparison of height, stride length and hand length, in surveillance footage such as this. (Photo: Courtesy of Essex County Prosecutor's Office)

Surveillance camera footage can be a crucial piece of 21st century evidence in crimes from robbery and kidnapping to homicide. But sometimes its quality, even at a digital level, can be grainy and not give a clear impression of a perpetrator.

But hand length, together with stride length and stature, could come together in a biometric triangulation helping to identify suspects, an international team reports in the latest Journal of Forensic and Legal Medicine.

“This triangulation between stature, hand, and stride/leg length provides a useful analysis for inferring and checking evidence from within a subject’s measurements,” they write. “Of interest in a forensic context may be the measurement, estimation, or inference of an individual’s height or stature. Indeed, it is a characteristic often reported on by witnesses or victims of a crime, and thus it has a real value in suspect apprehension.”

Prior investigations had connected two of three of the factors: between height and stature, and stature and hand length. But the new paper by scientists at the University of Kent, the University of Southampton and the University of Dundee is the first to bring together all three, they write.

Currently investigators can base height estimations off other reference points in the image or video; for instance, the comparative stature of an inanimate objects like a snack stand in a convenience store. But those visual clues may not always be available to forensic analysts. 

The scientists hooked 97 Caucasian subjects up to a Microsoft Kinect device. The imaging captured the movements—walking, stopping and turning—of the subjects, who were all aged between 18 and 35.

The device picked up 20 skeletal position points throughout the video feed—and calculated leg length and a height estimation.

The subjects also placed their hands on an acetate sheet with positioning pegs to obtain hand measurements.

The calculations then compared the self-reported height of the men and women, with their gait, their hand length and the overall calculations comparing all three. Together, they were consistent and accurate, within a margin of error.

“The significant correlation demonstrates the consistency of the measures extracted from the Kinect in relation to actual reported measures,” they write. “It is possible to envisage the use of these models under investigative scenarios where one or more facet of identity is available and a test is required on an actual or inferred measurement from the same subject. In a simple case, the modelled properties of stature may be required given a particular known hand measurement (or vice versa).”

Height and size estimations have improved as the ubiquity of surveillance cameras has given it increasing forensic investigative use. For instance, a biometric method by a Korean team published two years ago determined they could estimate height of people on footage from overhead cameras to within about a single centimeter.