Snapshot 3D image projection using light programming

LOS ANGELES - June 13, 2026 - PRLog -- Researchers at the UCLA Samueli School of Engineering and CNSI (California NanoSystems Institute), led by Professor Aydogan Ozcan, introduced a snapshot 3D image projection system that integrates a digital encoder with a passive diffractive optical decoder, jointly optimized end-to-end through deep learning. The hybrid architecture projects multiple, distinct images onto closely spaced axial planes in a single shot, marking a significant step toward compact, high-fidelity volumetric display technologies.

3D image display technology is essential for next-generation holography, immersive visualization, and augmented and virtual reality (AR/VR) interfaces, where accurate focal cues across depth are critical for natural depth perception and visual comfort. However, dense depth multiplexing in conventional holographic displays remains a challenge: as the axial image planes approach one another in the output volume, diffraction-induced cross-talk rapidly degrades depth selectivity and image fidelity.

The approach developed at UCLA addresses this challenge by pairing a learned digital encoder with a passive multi-layer diffractive decoder composed of structurally optimized surfaces. The encoder, built around a Fourier-based neural network, extracts multi-scale spatial and frequency-domain features from the target image stack, incorporates axial position information, and produces a single phase pattern that simultaneously represents all images to be projected in 3D. The encoded wavefront then propagates through the structurally optimized diffractive surfaces, which physically perform depth-dependent field programming during light propagation, optically routing image content to its designated axial depth while intrinsically suppressing inter-plane leakage.

Through numerical simulations, the researchers demonstrated multi-plane snapshot image projection with axial plane separations on the order of a single wavelength and showed that the system scales to volumetric scenes containing 28 axial slices encoded into a single phase pattern. The work also characterized key design factors, including the depth of the diffractive decoder, output diffraction efficiency, spatial light modulator resolution, and axial encoding density, providing practical design guidelines for future diffractive 3D displays.

The team further experimentally validated the framework using a two-plane optical prototype with a single-layer physical decoder operating in the visible spectrum. The measured intensity patterns closely matched both the numerical simulations and the target images, and clearly outperformed a free-space baseline without a diffractive decoder, confirming the feasibility of the hybrid digital-optical architecture for snapshot 3D image projection.

Looking ahead, the framework could be extended to multispectral operation, multi-perspective holography, and physically fabricated multi-layer passive decoders for compact, energy-efficient 3D display systems.

The study was conducted by Dr. Çağatay Işıl, Alexander Chen, Yuhang Li, F. Onuralp Ardic, Dr. Shiqi Chen, Che-Yung Shen, and Prof. Aydogan Ozcan of the UCLA Electrical and Computer Engineering Department and the California NanoSystems Institute (CNSI) at UCLA.

Article: https://doi.org/10.1038/s41377-026-02378-3