Russian biometrics firm Vocord has revealed that its identification algorithm got the highest score in a recent MegaFace global competition, correctly recognizing more than 75% of the faces in the MegaFace database – the highest score worldwide.Vocrod develops intelligent video surveillance and audio logging systems and the 3D facial recognition system VOCORD FaceControl 3D, and 2D facial recognition solutions.The internationally recognized MegaFace online testing platform is a database consisting of more than one million photos in which more than 690 thousand different individuals are pictured, shot in “natural” conditions, i.e., in random poses, with free facial expressions, lighting and exposure. The goal of the competition is to evaluate the accuracy of facial recognition algorithms from the gallery sample against an increasing number of “distracting faces” (from 10 to 1 million) from the MegaFace database. The MegaFace competition allows for the evaluation of existing facial recognition algorithms on a large number of images.The VOCORD recognition algorithm, based on deep neural networks, demonstrated superior results in world-wide testing in recognition of grade 1 faces, the company revealed in a statement.The tests demonstrated that the VOCORD algorithm correctly recognized 75.127% of the faces, which the firm says is the best result in in the world at the present time.In describing the success of VOCORD's algorithm, company Technical Director Alexey Kadeyshvili said “The excellent facial recognition results achieved by VOCORD specialists are the result of systematic, painstaking scientific work which we have conducted over the course of many years. At present, the company also continues its efforts in optimization of developed algorithms for facial verification tasks. In particular, we have very interesting results in combating fraud in facial authorization”.
Select Page















