CyberExtruder released its new Aureus 3D facial reconstruction software, which has shown an increase in facial recognition matching accuracy by 26.5 percent when compared to even the best 2D facial recognition products.
The National Institute of Standards and Technology (NIST) first recognized the value of utilizing 3D information as a biometric tool in the Facial Recognition Vendor’s Test (FRVT) of 2002 and in FRVT of 2006. The NIST reported findings of FRVT 2006 and the Iris Challenge Evaluation 2006 established the first independent performance benchmark for both 3D facial recognition and iris recognition technology, which concluded that the performance of both is comparable.
Despite previous technology advancements, law enforcement and security professionals have continued to struggle with a lack of availability of quality images with which to identify and match “persons of interest.” Most surveillance cameras are mounted overhead so they can follow a person’s movements through a scene. Because of this, individual frames taken from these cameras make it difficult, if not impossible, to identify the person unless they happen to look directly into the camera.
Software constructs 3D models from 2D images
Aureus 3D significantly improves the accuracy of facial recognition systems by taking a subject’s facial image and reconstructing it into a 3D format that factors in varied poses and facial expressions, and compensates for poor lighting and partial images. When Aureus 3D reconstructs a face, its patented processes are able to determine the person’s original pose relative to the camera. This information enables Aureus 3D to produce results which lead to matches when a subject may be turned away from the camera as much as 70 degrees and a downward (or upward) angle of 25 degrees.
Better models make better matches
The product’s resulting 3D model, called a 3D facial template, is used to perform identification and verification tasks.