May 25, 2021

Visualization of the optimized phase delays, the output of the model and the ideal output of holograms.

Imagine being able to see a hologram of the person you’re talking to on your cell phone. Sound futuristic? Mizzou Engineers are finding ways to someday make that a reality. A research team is studying how artificial intelligence (AI) can help generate holograms in everyday settings. Marshall Lindsay, a PhD student in computer science, presented the work at an international conference hosted by SPIE, the Society of Photo-optical Instrumentation Engineers.

Unlike traditional holograms, which are made by using laser beams to illuminate an object and intercepting it on film, Lindsay’s work digitally computes a filter to modify an image or object to become three-dimensional.

“So it might be a filter that makes it looks as though your nose is coming out of the screen or makes your 2D image look like a 3D image,” said Scott Kovaleski, a professor in the electrical engineering and computer science department and co-author of the paper.

For the paper, Lindsay demonstrated a digital representation of a hologram using the Mizzou Tigers logo. Now, the team is determining how to manipulate light waves so that if a laser were projected through it, it would create a 3D image of the logo.

Once researchers have taught the AI to predict and create a hologram, eventually it could be developed in the physical world. Kovaleski compared it to the 3D images that appear in newspapers in the world of Harry Potter.

“If we had an ‘ink’ that could realize those changes in light, we could print the Harry Potter newspapers,” he said. “With a pattern of ‘ink,’ we would know how to lay down a layer on the newspaper, and when the light reflects off the ink, we would get an image of a character in depth and perspective. We aren’t making the ‘inks,’ but rather are finding the patterns that, when printed, would produce holographic images.”

And Lindsay believes it’s just a matter of time before that means we could use holograms in new ways.

“We have to get smaller and faster, but we could have a piece of material you put over the flashlight of your phone’s camera, and you see your grandma,” he said. “I 100% believe that’s going to be a thing.”

Associate Professor Derek Anderson and Charlie Veal, a PhD student in computer science, were co-authors of the paper, “Machine Learning Assisted Holography.