Communications of the ACM
What Makes Tom Hanks Look Like Tom Hanks?
University of Washington researchers have demonstrated that machine learning algorithms can capture the persona and create a digital model of a well-photographed person like Tom Hanks from the vast number of images of them available on the Internet.
Credit: University of Washington
Researchers at the University of Washington (UW) have developed a system that uses machine-learning algorithms to construct and animate three-dimensional (3D) models of celebrities based solely on large numbers of photographs.
Current 3D modeling technology, such as that used by major film studios, requires a painstaking process to capture every angle of a person's face and movements. The system designed by the UW researchers is meant to be much simpler. It combines several techniques developed over the last five years by a research group led by UW professor Ira Kemelmacher. The techniques include advances in 3D face reconstruction, tracking, alignment, multi-texture modeling, and puppeteering.
The system works by feeding at least 200 images of an individual, taken over time in various scenarios and poses, to machine-learning algorithms. The raw data is then used to create the 3D model, which can then be animated so it mimics a given person's distinctive traits.
The researchers say the technology could some day be used to create 3D models of absent loved ones or celebrities that people could interact with as though they were real.
From University of Washington News and Information
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA
No entries found