An Identification System that Distinguishes between Identical Twins

Michael

Alex

The identical twins Michael and Alex Bronstein, who are graduates of the Program of Excellence, were able to build a face identification system, which can distinguish even between identical twins, a phenomenon that the human eye finds difficult to do. The system has great potential for application in security facilities, which currently suffer from low reliability due to their sensitivity to lighting conditions and changes in head position and facial expressions.

According to researchers, the central problem in a three-dimensional face identification is finding the similarities between the facial surfaces. The existing methods have failed since they compared facial surfaces as rigid unchanging surfaces. The system devised by the Bronstein brothers and Prof. Ron Kimmel from the Department of Computer Sciences at the Technion, is based on creating a three-dimensional image and comparing the geometrical variables of the face as flexible surfaces.

The heart of the 3D face recognition algorithm is based on a prior work of Asi Elad and Prof. Kimmel, who proposed an elegant way to compare isometric surfaces. The main idea was transforming the surface to a "canonized form" by a process known as isometric embedding, in which the distances between the various points on the surface are maintained. The resulting canonical form it is not sensitive to changes in head orientation, lighting or facial expressions.

Michael Bronstein about the project: "The heart of the 3D face recognition algorithm is based on a prior work of Asi Elad and Prof. Kimmel, who proposed an elegant way to compare isometric surfaces. The main idea was transforming the surface to a "canonized form" by a process known as isometric embedding".