The human face is inherently a 3D object. However, current facial recognition techniques utilize 2D images to match surveillance photos with a database of “enrolled” photos. As long as the two photos have relatively similar pose, lighting, and expression, matching results can be acceptable.
However, when surveillance photos are compared with a database of frontal enrollment pictures, matching performance drops dramatically and positive identification fails.
Genex has developed a breakthrough technology that harnesses the power of 3D imaging to create multiple 2D images that accurately present the face in varying pose, lighting, and expression conditions. 3D Enroll allows seamless integration with existing 2D recognition databases.
Now users can deploy a 3D recognition solution while maintaining a link to legacy systems.
Features:
-
Pose – Generates accurate multiple poses of an individual in both pan and tilt.
-
Lighting – Generates a rich variety of lighting conditions with multiple light sources.
-
Expression – Generates common expressions by adjusting muscles in the face.
Product Brochure: Genex 3DEnroll (requires PDF reader) |
