Social services and the police are now searching for the missing. But perhaps, as reported by The New Atlas, soon machines will come to the aid of a person, which, using complex algorithms, will find people, even if they disappeared several years ago.
The neural network created by scientists matches human facial features, such as lips, forehead, cheek shape and others. Based on the data received, the system begins to analyze how a person could have changed over time. In addition, the program even takes into account the external data of relatives, individual signs of aging and much more. This radically distinguishes the program from all existing analogues.
The researchers emphasize that their algorithm can be used both to find missing people and to find criminals. As an example, experts used a photo of 21-month-old baby Ben Needham, who disappeared in 1991 on the Greek island of Kos. Although the baby was never found, the system gave a rough idea of what Needham might have looked like at the age of 20 and older. Project developer Ali Maina Bukar said
“We train the system by feeding it hundreds of thousands of facial images with appropriate ages. After that, seeing the aging process of a person, the neural network generates images on the fly. This is a very fast way. All existing algorithms today use linear or one-dimensional methods, while our neural network is non-linear, which means it is better suited for each specific person."
VLADIMIR KUZNETSOV