Researchers from Universidad Carlos III de Madrid (UC3M) have identified a new approach to facial biometrics that yields 95% accuracy, according to a Science Daily article. As opposed to other facial biometrics systems, which utilize predetermined characteristics of the face to authenticate a user’s identity, the new system takes in the whole face and determines which aspects are the most telling of the individual’s identity and authenticates via those traits.
One of the additions to basic facial biometrics systems, which utilize facial symmetry and facial trait geography they made, was including facial texture. The real capability, according to the researchers, was found in how they combined the facial symmetry function with the facial texture.
Despite the advantages of their system over other facial biometric systems, they also acknowledge certain short failings of the system including its dependence on consistent lighting and failures due to aging effects on the face.
Read the full story here.