Facial recognition challenge

Facial recognition allows biometric verification in all situations. It is the subject of much research for safety-related applications (criminal identification, verification of traveler identity, security control in companies). For commercial applications, facial recognition should enable secure online transactions.

It is based on automatic feature recognition of faces (distance between the eyes, the edges of the nose, corners of the mouth, ears, chin, etc.) compared with those registered in advance in a database. The development of 3D sensors, of softwares associated with the study of moving faces, profile views, the aging models still opens the hope of many improvements.

The effectiveness of the facial recognition depends on several key factors, including the quality of the image, identification algorithms and the availability of reliable databases. The most interesting application seems to be the buyer authentication in online shopping, instead of communicating a password. But there is still some reluctance to overcome, notably that of protection of personal privacy. Nevertheless, the Russian site FindFace offers the opportunity to find a person on the Internet using a single photograph taken with a smartphone. A basic 300 million photos of data would achieve a similar accuracy of 70%, say the managers of the Russian social network Vkontakte and of N-Tech.Lab creator. China is no exception in this area, but international copyright rules forbid us to say more about the revolutionary application that is being developed for online shopping. See also www.morpho.com/  https://fr.sputniknews.com/

© 2015 Strategies Telecoms & Multimedia | Contact |  -