Nottingham scientists have created a form of artificial intelligence which can build a ‘3D selfie’ from a single photograph.
Computer scientists at the University of Nottingham have developed technology capable of producing 3D facial reconstruction from a single 2D image in a new web app – solving one of the trickiest problems in computer graphics.
More than 400,000 people have so far tested it by uploading their own selfies on the research website, which creates the images in seconds.
The research was led by PhD student Aaron Jackson and carried out with fellow PhD student Adrian Bulat in the School of Computer Science.
It used a type of artificial intelligence which gives computers the ability to learn things without programming – meaning just one photo is needed to create a 3D picture instead of multiple images.
The research team, supervised by Dr Yorgos Tzimiropoulos, trained a system with a huge database of pictures so it could learn to build images.
Dr Tzimiropoulos said: “The main novelty is in the simplicity of our approach which bypasses the complex pipelines typically used by other techniques.
“We instead came up with the idea of training a big neural network on 80,000 faces to directly learn to output the 3D facial geometry from a single 2D image.”
Apart from face and emotion recognition, the technology could be used to personalise computer games, improve augmented reality, and let people try on online accessories such as glasses.
It could also have medical applications including simulating the results of plastic surgery or helping to understand medical conditions such as autism and depression.
The results will be presented at the International Conference on Computer Vision (ICCV) 2017 in Venice next month.
