Person identification by gait analysis and photogrammetry.

Surveillance images from a bank robbery were analyzed and compared with images of a suspect. Based on general bodily features, gait and anthropometric measurements, we were able to conclude that one of the perpetrators showed strong resemblance to the suspect. Both exhibited a gait characterized by hyperextension of the leg joints, and bodily measurements did not differ by more than 6 mm on average. The latter was quantified by photogrammetry: i.e., measuring by using images of the perpetrator as captured by surveillance cameras. Using the computer software Photomodeler Pro, synchronous images from different cameras were compared and concurrent body features were identified. The program could then render the perpetrator as a three dimensional, high-precision, scalable and measurable object.