Cognitech, based in Pasadena, Calif., and a DME pioneer, is best known for its “Forensic Video Tri- Suite” software that contains advanced video analysis and allows the viewing of multiple-video streams from different platforms within one application. The platform's video demultiplexing function allows the software to better sort video according to content, the company said.
Cognitech is on the verge of releasing an upgrade to its software platform called “Video Investigator 13,” which will include “Face Fusion 3D,” a first-of- a-kind application to enhance and super-resolve images of human faces from low-resolution surveillance videos. The software will clean up video of subjects walking and moving around, while their faces and heads are seen from very different angles, said Lenny Rudin, the company's co-founder.
It is unique in the forensic-video field because it utilizes the 3D shape of human faces in order to reconstruct better facial shots, which are then used for identification and comparison, Rudin said. Previous pixel-based forensic-image enhancement software does not use 3D-shape geometry, Rudin said.
“This new capability is the game changer when it comes to enhancing, rotating and generally moving images of complicated 3D shapes like human faces.”
Future software updates will allow for the enhancement of images of other body parts such as an arm or leg with tattoos rapped around them, Rudin said.
FBI image is the publicly released image that is enhanced as compared to original video, and released by FBI from a blurry original frame from the same public video. We naturally assume that that was their best shot, as at the time the suspect was not apprehended, and they appealed to public help.
Cognitech image is a result we get with Cognitech's 3D Frame Fusion when applied to the same above mentioned frame of the Boston Marathon Bombing Surveillance Video. The improved facial features definition over the FBI released public image is obvious.