The Chinese manufacturer purchased the patents and hired the team of designers of Vocord Company, resident of the IT cluster of the Skolkovo Foundation. The deal was confirmed to Sk.ru by Vocord.
The parties will not disclose the price of the deal. The sources in Vedomosti Daily assess the deal as being worth around USD 40-50 million.
Vocord was founded in 1999. The company develops cameras and software for face recognition, video monitoring and tracking of traffic rules violations. At Skolkovo, where the company became a resident in 2011, the team has been working on the project of VOCORD FaceControl 3D, the 3D system for non-cooperative face recognition. Unlike most other IT cluster projects that mostly have to do with programming, this system combines hardware and software elements. Huawei, experts believe, could be interested in Vocord’s HD cameras with smart functions.
Vocord has won many medals at MegaFace, the world championship of face recognition. During one of the events the Russian company’s algorithm proved to be more effective than that of Google. Sk.ru has written about the successes of this Skolkovo team in commercializing research results on many occasions. Computer systems recognize people’s faces better than other people do. We can only recognize faces of familiar people – family members, friends, neighbors, etc. If someone should get a stack of 1000 photographs and ask to find one photo analogous to a certain other photo in the database, they would run into a problem. Naturally, specialists do much better than untrained people, but machines have been performing better than people in recognition of faces of strangers for quite a few years now. The main problem with recognizing faces in public places, such as, for instance, transport hubs, is that people tend to refuse to “cooperate.” The passengers smile, frown, talk, eat and often do not look straight at the camera. All these factors considerably lower the reliability of traditional systems of face recognition.
Unlike some other biometric systems that require the face of the person to be positioned precisely in relation to the camera, the VOCORD FaceControl system can work in the fully non-cooperative mode. This means that the person does not need to position him- or herself in front of the camera in a certain way: the system focuses on the face of a particular person in the flow of other people who may not even be aware that they are in the vision field of the system. The system can be used on the underground, on railway stations, in stadiums, cinema theatres, stores, etc. Face recognition systems are in demand and can be widely used, for instance, in retail trade, where the VOCORD FaceControl system helps deal with two major groups of situations: safety (helps lower the risks of shoplifting and investigate various incidents), and marketing (helps improve service quality, define target audiences and optimize advertising costs).