UCLA team shows one-shot 3D projection method for 28 image layers

Researchers at UCLA have developed a 3D image projection method that can place multiple images at different depths from a single programmed light pattern. The work points to a possible route for smaller volumetric displays used in holography, augmented reality and other depth-sensitive imaging systems.

The UCLA Samueli School of Engineering and the California NanoSystems Institute said the system combines a digital encoder with a passive optical decoder made from diffractive layers. The study, led by Professor Aydogan Ozcan, was published in Light: Science & Applications.

Conventional holographic displays can struggle when image planes are packed closely together along the depth axis, according to the UCLA team. As those planes get nearer, diffraction can cause light meant for one depth to spill into another, reducing image clarity and depth separation.

Digital encoding paired with optical decoding

The method starts with a digital encoder built around a Fourier-based neural network, according to UCLA. That encoder analyzes the target stack of images, including spatial features, frequency-domain information and the intended axial position of each layer.

From that information, the encoder produces one phase pattern that contains the instructions for the full 3D image stack. Once light carrying that pattern passes through the passive diffractive decoder, the structured surfaces alter the field during propagation and help send image content to the assigned depth planes, the researchers said.

UCLA described the design as a hybrid digital-optical system: computation prepares the light field, while the passive optical elements perform depth-dependent routing without an active display element at each plane. The researchers said the decoder also helps reduce leakage between planes.

Simulations reached 28 axial slices

In numerical tests, the team reported snapshot projection across planes separated by distances on the order of a single wavelength. The simulations also showed the approach scaling to volumetric scenes with 28 axial slices encoded into one phase pattern, according to the published study.

The researchers also examined design limits and tradeoffs, including decoder depth, diffraction efficiency, spatial light modulator resolution and the density of axial encoding. UCLA said those results provide design guidance for future diffractive 3D display systems.

The team then tested the concept with an optical prototype operating in the visible spectrum. That experiment used two image planes and a single-layer physical decoder, rather than the larger multilayer structures explored in simulation.

According to UCLA, the measured intensity patterns from the prototype closely agreed with simulations and the intended target images. The researchers also reported that the prototype performed better than a free-space comparison setup that did not include a diffractive decoder.

Possible display and imaging uses

The UCLA team said the approach could support high-axial-resolution 3D projection in compact systems. Possible uses identified by the researchers include holographic and near-eye AR/VR displays, multi-depth volumetric microscopy, real-time 3D visualization and volumetric optical computing.

The study lists Çağatay Işıl, Alexander Chen, Yuhang Li, F. Onuralp Ardic, Shiqi Chen, Che-Yung Shen and Aydogan Ozcan as authors. UCLA said future versions could extend the framework to multispectral operation, multiperspective holography and physically fabricated multilayer passive decoders.

This story draws on original reporting from Phys.org.