Through the program, the scientist are capable to reconstruct a person’s face and facial expressions down to the last details.
Ever wondered what it will be like to actually talk to someone and that person to be hundreds of miles away? This idea has been highly debated when it came to long-distance communication system involving holographic projections. What if there’s no need for a hologram when it comes to person-to-person long-distance video transmissions? Virtual celebrities lookalikes are the new fad but from a specific point of view.
Recently, a team of computer engineers hailing from the Washington University have developed a technique which could very well prove to game-changing when it comes to augmented virtual reality. The team starting their project from a rather bizarre question: what makes Tom Hanks to be Tom Hanks? Apart from being a very gifted actor, Tom Hanks has also made it to a lot of magazines due to his unique facial expressions.
Taking that into account, the scientists managed to develop a computer program which is able to mimic a person’s facial expression down to the last details. In terms of puppeteering, we had seen some very vivid examples were a ventriloquist would change the puppet’s tone in an instant in order to create a comical effect. With this technology, we are capable of doing just the same thing, but at a whole new level.
Image if you could take the figure of a famous actor, let’s say Bruce Wills and made his voice sound like Chris Tucker’s? Or if former president George W.Bush could talk like Ian McKellen? The possibilities of this novel computer program are quite infinite. Basically, by using this unique facial recognition program, a film producer will be capable of making a whole movie without using any real-life actors. All he has to do is feed the program different pictures of the actors, depicted any number of facial expressions.
In fact, that was what the team of researchers did in order to come up with these outstanding results. For them to perfectly recreate Tom Hank’s persona, so to say, the team searched high and low through Google in order to find photographs with the actor. More than 200 pictures were fed to the machine in order to sketch up a portrait. The result was that the program was able to recreate the facial expressions of the actor without any margin for error.
Professor Ira Kemelmacher-Shlizerman, the think-tank behind this amazing project said that he and his team were able to recreate Hank’s persona by using a set of facial recognition algorithms. The whole contraption uses any number of technological marvels such as 3D facial reconstruction, tracking sensors, alignment algorithms, multi-texture modeling and some techniques borrowed from the art of puppeteering.
Photo credits:www.wikipedia.org
