Researchers at the University of Bradford have developed a new face-aging method that teaches the machines how we, humans age. This is done by feeding the algorithm facial feature information from a giant database of people at various ages. The new technique improves existing techniques, achieves a greater level of accuracy and is expected to boost the search for missing people.
As Professor Hassan Ugail of Bradford’s Centre for Visual Computing who is leading the research explains, each year around 300,000 people disappear in the UK alone. That was the motivation behind the development of the new technique.
The new method can map out the key facial features, including the shape of the mouth, nose, forehead, and cheek at a certain age. This is then fed to a computer algorithm which synthesizes new, older-looking features for the specific individual at different ages.
Using predictive modeling, the algorithm can predict age progression for each person. This is done by incorporating facial information from a database of people at different ages. The researchers then test the results through a method called de-aging, which is when they take an individual’s picture and run the algorithm backward. This is then compared to the actual photo of that individual, usually taken at the young age.
Source:
University of Bradford via Phys.org (https://phys.org/news/2017-06-face-aging-technique-boost-people.html)
The Wonderful World of Algorithms 