Many children in the 1980s will remember the many 3D video games that will appear in the malls and arcades. Many of us thought the future would be full of 3D games and movies, but that didn’t happen. MIT has developed a new method for generating holograms using artificial intelligence to create 3D content in real time. The new method is called 3D imaging and it can be run on a smartphone or laptop.
Massachusetts Institute of Technology Researchers Say current VR headsets can make users feel sick because virtual reality simulates 3D rendering on a static 2D screen. Researchers believe that the best way to display 3D content could be reconstructed 3D images of the digital world. Stereographs can provide excellent representations of the 3D world by changing perspective based on the position of the viewer. This fact allows the eye to adjust the depth of focus to focus alternately on the foreground and background.
Researchers were working on creating computer-generated holograms, but that process required a supercomputer to run massive amounts of physical simulations. Not only was it a waste of time, it produced less than realistic results. The new MIT method can produce stereoscopic images almost instantly. It uses a learning-based methodology that is in-depth and efficient enough to run on a laptop or smartphone very quickly.
Tense holography can finally meet the 10-year goal of commercially available holography, says MIT researcher Liang Shi, the lead investigator on the project. Xi also believes that advances could bring holography to areas such as virtual reality and 3D printing. Holograms encode the brightness and phase of each light wave together to provide a more accurate depiction of scenery and depth of scene.
Computer generated holograms require that each point in the scene have a different depth, so the same operations cannot be applied to every scene, which greatly increases the complexity. Traditional holograms require a supercomputer assembled to run physics-based simulations and take seconds or minutes to create a single hologram. Team deep learning methods accelerate computer generated holograms allowing real-time 3D image creation.