Virtual reality is becoming more and more realistic, and in three years we won't know if we're talking to a person or an avatar
Scientists have presented in the USA the most realistic virtual reality room possible, where XNUMXD characters can be created in extremely high resolutions.
The cube-shaped room, each side of which is 3 meters long, is called C6 and it was built at the Center for Virtual Reality Applications at the University of Iowa, with an investment of 5 million dollars in equipment. The resolution of the projection in the room is more than double that achieved in any other virtual reality room in the world, the researchers say, and 16 times more pixels are available in the room today compared to its previous version built in 2000.
Among the equipment is a cluster of HP computers, containing 96 graphic processors, 24 Sony digital projectors and an eight-channel sound system, as well as infrasound motion tracking technology.
The project is supported by the US Department of Defense for the Office of Scientific Research of the Air Force. The facility will be open to the public with pre-registration at the emerging technologies conference that will be held at the university on April 25-27.
At the same time, the University of Illinois in Chicago and the University of Central Florida in Orlando have developed a system that will allow us to talk in XNUMXD with Einstein or Newton, in such a realistic way that it is possible to think that the images are of living people who are in the room with us.
Technologies used in computer games, animations and artificial intelligence are used as components of the project, which received half a million dollars of support from the National Science Foundation, for three years with the aim of combining these components together to create a methodology that will make virtual characters common.
The University of Illinois will focus on computer graphics and interactions, while the University of Central Florida will focus on royal intelligence and natural language processing.
"The goal is to combine artificial intelligence with the latest developments in the field of graphics and video games, to allow us to establish a historical database of people that will be much more realistic than those we obtain today through text, voice, and video clips" says Jason Leigh (Leigh), professor of science The developer and director of the Electronic Visualization Laboratory at the University of Illinois.
The lab will build special studios that will be used to capture movements, to digitize the images and movements of real people. The knowledge will be stored in databases, the voices will be analyzed to create them again in a synthetic way that will sound natural. The gestures of the body will also be taught in the way we know technically as avatars.
Lei and his team hope to create virtual humans that can respond with a high level of recognition to different voices and multiple ways of asking the same questions.
"Imagine a computer powerful enough to contain an avatar that responds "do you understand what I'm saying" in a natural way that humans communicate with each other" says Leigh. "We're trying to jump ahead and make the computer as natural as possible." According to him, the XNUMXD visualization techniques used today in computer games and in virtual environments (such as Second Life) were born in laboratories such as the one in Illinois, and even today game manufacturers are trying the next generation techniques in his laboratory and in other laboratories.
One response
The response was received from Eyal
Things sound very nonsensical to me, are you claiming that in 3 years a computer will be able to successfully pass the famous Turing test?? This is the holy grail of the world of computing and robotics! This requires a computer with true artificial intelligence, a subject that is still in its infancy! How does this fit with what is written here? Also, could you give a link to a video or photos that demonstrate witness