If we thought that selfie fever could not go further, we were wrong. The 3D selfie comes to revolutionize –even more– the complex world of social networks, and what we colloquially know as postureo.
How do you get a 3D selfie? Artificial intelligence has the answer. It is a technology developed by computer scientists of the University of Nottingham and Kingston University that, through an artificial intelligence system, is capable of reconstruct all 3D facial geometry from a single 2D image, including non-visible parts of the face. This technology, although not perfect, represents a great advance in this field.
The technique is based on a complex process starring the artificial intelligence; we speak of a new technology based on the convolutional neural network, capable of using machine learning so that computers can learn without being programmed for it. The work of this mechanism is to transform conventional 2D images into a 3D selfie, something that can only be achieved through a reconstruction of the 3D facial geometry.
But the key that makes this project novel lies in introduce the way of learning our brain in an app, that is, a process based on the advancement of deep learning, a form of mechanical learning that uses artificial neural networks to mimic the way the brain makes connections between pieces of information.
Reconstruct 3D facial geometry
The research team, supervised by the dr. Georgios Tzimiropoulos, he was able to develop that convolutional neural network based on a huge data set of 2D images and 3D facial models; with all this information, his neural network was able to reconstruct the 3D facial geometry from a single 2D image, including non-visible parts of the face by making a series of conjectures.
Through a web application, anyone can upload a color image and receive, in a few seconds, a 3D model of it. More than 400,000 users have already done the test. In addition, this technology could mean great advances in the world of security, video games and virtual reality. We will follow his next steps closely.
