Researchers at the University of Nottingham and Kingston University in the U.K. have developed a new Web application that can render a two-dimensional (2D) facial image as a three-dimensional (3D) construct.
The researchers say people can use the app to upload a single color image and quickly produce a 3D model showing the shape of their face.
The team trained a machine-learning convolutional neural network on a massive dataset of 2D pictures and 3D facial models, enabling it to reconstruct 3D facial geometry from a single 2D image while also deducing the non-visible parts of the face.
"The main novelty is in the simplicity of our approach, which bypasses the complex pipelines typically used by other techniques," says Nottingham professor Georgios Tzimiropoulos. "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."
From University of Nottingham (United Kingdom)
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA