Researchers at the Massachusetts Institute of Technology (MIT) announced on Wednesday that they had developed a technology capable of producing 3D holograms in real time. The method is based on artificial intelligence and can be performed even on notebooks.
The approach is called “tensor holography”, something like “tensor holography”, in Portuguese. According to scientists, the novelty may boost fields such as virtual reality and 3D printing.
The lead author of the study was Liang Shi, a doctoral student in the Department of Electrical Engineering and Computer Science at MIT. He said that, contrary to what people say, the technology the world currently has access to can perform 3D holography calculations in real time.
“It is often said that holographic screens will only be commercially available in 10 years. The problem is that this statement is decades old ”, he argued. The research was published yesterday in Nature, one of the most important scientific journals in the world.
While a photograph encodes the brightness of each light wave, a hologram is able to encode the brightness of each phase of each light wave. With this, the technology is able to represent a figure with depth and not just as flat as the photo.
The innovation developed at MIT uses deep learning, which is a branch of machine learning, to accelerate the production of a hologram by the computer. To support the technique, a customized database of 4,000 pairs of images was built.
To create the holograms, the scientists gathered all the color and depth information from the image pairs that are processed from physics calculations. The learning process of computers and the speed at which calculations were performed to create the figures surprised the experts.
In addition to the possibility of using technology in more popular mechanisms such as virtual reality, the solution offers benefits such as more “realism and comfort for the observer”, emphasizes the work. The research was supported by companies like Sony.
Leave a Reply