Researchers recreate Mona Lisa using 'ground-breaking' holography technology

Researchers use deep learning to create lifelike 3D images

A holography technique developed by researchers in China and the United States can recreate the Mona Lisa in stunning detail. Photo: Nicolas Tucat

Researchers have recreated the Mona Lisa using a new holography technique.

This work was conducted by an international team of researchers from China and the United States, and their findings were published in Applied Physics Reviews.

The researchers used metasurface-based holography, a technique involving a deep-learning algorithm, to generate and refine a hologram of the Mona Lisa.

Holograms are images created by capturing and reconstructing the interference pattern of light or sound waves, used in fields such as entertainment, medical imaging and communication.

According to author Yue-Sheng Wang, metasurfaces are two-dimensional materials composed of tiny, antenna-like structures that help control the phase and amplitude of the waves interacting with the metasurface.

This control aids in the creation of the holographic image.

The team developed a metasurface optimisation method to improve the holographic process.

They used a deep neural network-based algorithm to adjust the antenna-like structures within the metasurface. By reducing discrepancies between the original and holographic image iteratively, they refined the metasurface, improving the hologram's quality.

The researchers used the Mona Lisa as a demonstration of their method's effectiveness, due to its universally recognised features and intricate layer transitions.

The method successfully reconstructed the Mona Lisa, including her left eye in significant detail.

The technique currently produces two-dimensional holograms, but the researchers suggest it could also be used for three-dimensional images.

The team's method allows precise control of sound waves, which they argue is vital for advancing non-invasive medical therapies, effective noise control, and optimising acoustic environments.

They believe that their technique could have significant implications for the field of holography and plan to improve its versatility, compatibility with 3D printing, and reduce the training time required.