Algorithms help create 3D animations of celebrities

Researchers at the University of Washington (UW - USA) recently demonstrated 3D animations by actors Tom Hanks, Daniel Craig and some other celebrities that can be created by concatenating stills and voices. oral pictures taken from the internet. Moreover, they showed how their algorithms can make digital images come alive, by creating faces that are saying things that are actually being told by another person.

All of the interesting achievements mentioned above are formed by 3D face reconstruction technology , tracking, alignment, layout simulation; which was developed by the research team led by Ira Kemelmacher-Shlizerman - assistant professor of computer science and engineering, working at UW. The latest progress of the research team includes the ability to convey ways of expressing a person's speech, to another individual. For example, in the demo video of scientists: "apply" the style of former US President - George W. Bush to the face of politicians and other celebrities.

All of these interesting achievements are formed by 3D face reconstruction technology.

It is known that the algorithm developed by Washington University researchers exploits at least 200 photos taken from the internet, in different situations. That process is called "learning from nature". Maybe some people will think of bad intentions, which can be easily done by this new technology. However, according to the researchers, their original goal was not "unrighteous" as we thought. Experts say they created this method in the hope of allowing family members to interact with their relatives who cannot always meet, by creating a thorough model. Their three-dimensional algorithm, from family photo albums, videos or other content.

As virtual reality technology is growing, scientists are probably heading for a 3D solution that will replace existing 2D communication facilities such as Skype or FaceTime. "Maybe one day, just put on the virtual reality glasses and you can see your beloved mother sitting on the familiar armchair ," Kemelmacher-Shlizerman said.

Mrs. Kemelmacher-Shlizerman - who leads the research team at the University of Washington.(Photo: gannett-cdn).

Besides, researchers at UW also claim that their technology can be applied to the entertainment industry, thanks to the replacement of the process of creating digital characters in the film. For example, to create a fictional character like Benjamin Button (in The Curious Case of Benjamin Button ), every angle and movement of actor Brad Pitt must be recorded in a controlled manner. The images are then edited and stitched together to create the complete character.

This week, the team is expected to present their project in detail at the International Computer Vision (Computer Vision) conference held in Chile.