Robber's Personal Identification by Morphometric Comparison Between Recorded Images and 3D Avatar of the Suspect

Personal identification based on 3D digital photogrammetry presents a natural evolution of a previous research in this field by "parameterized superimposition" (PS). The authors' experience was based on 2D/2D comparison between video frames taken from the surveillance camera system during the robbery and frames of the suspect brought back in the same place where the robbery was perpetrated. This technique involves four steps: The PREPARATORY PHASE, in which the recorded images of the robber are studied and improved. Frames with better view of robber's face landmarks are then chosen. The 3D ACQUISITION PHASE, during which a 3D photogrammetric avatar of the suspect face is created; this phase only requires 4 photos made simultaneously with a calibrated camera. The SUPERIMPOSITION PHASE is preparatory for the final step and involves a meticulous spatial orientation of the 3D avatar in the same position taken by the offender in the selected frames. A snapshot of the 3D avatar is now taken. During the METRIC IMAGE ANALYSIS a quantitative comparison between the image of the robber's face and the snapshot obtained is used. To perform this step it is necessary to clearly recognize at least five landmarks on the robber's face using a suitable software.