Computational Imaging with Metasurface Technologies

Computational imaging fuses physical wavefront modulation with algorithmic reconstruction to surpass the limitations of conventional optical and microwave systems. Metasurfaces—two-dimensional assemblies of subwavelength elements—enable unprecedented control of amplitude, phase and polarisation of electromagnetic waves. By encoding scene information into customised field patterns and recovering high-resolution images through inverse-scattering solvers, this paradigm decouples hardware constraints from image quality. Recent advances have explored programmable elements, frequency-diverse apertures and semantic priors to tackle ill-posed inverse problems and to extract richer information such as depth, polarisation signatures or semantic content. Applications span human-scale millimetre-wave sensing, touchless security screening, medical diagnostics and smart manufacturing. The harmonised design of metasurface hardware and computational algorithms is key to realising compact, lightweight and cost-effective imagers that operate across the electromagnetic spectrum.

Research from Nature Portfolio

Recent studies have introduced semantic regularisation into the reconstruction of three-dimensional and four-dimensional compressive microwave images. By integrating language-based priors into inverse-scattering workflows, these approaches allow targeted concealment or transformation of selected subjects during reconstruction, advancing privacy protection and touchless interaction. Earlier foundational work demonstrated single-sensor, single-frequency microwave imaging using a two-bit programmable metasurface. By generating random binary transmission masks on a planar aperture, the system solved inverse-scattering problems at a fixed frequency, drastically reducing hardware complexity. Complementary efforts employed frequency-diverse metasurface panels swept across millimetre-wave bands to reconstruct diffraction-limited images of human-scale scenes. Key computational calibration methods enabled rapid, high-fidelity imaging despite electrically large apertures, paving the way for low-profile holographic sensors.

Computational Imaging with Metasurface Technologies publication trend

The graph below shows the total number of articles in computational imaging with metasurface technologies across all publications each year (not limited to Nature Index journals).

Technical terms

Metasurface: A planar assembly of subwavelength resonators that manipulates electromagnetic waves.

Computational imaging: A method that encodes scene information into structured measurements and recovers images via algorithmic reconstruction.

Inverse-scattering problem: The challenge of deducing object properties from measurements of scattered fields.

Frequency-diverse aperture: An imaging approach that sweeps or codes multiple frequencies to generate spatially varied illuminations.

Semantic regularisation: The incorporation of high-level, language-derived priors to guide the solution of ill-posed inverse problems.

Programmable metasurface: A metasurface whose individual elements can be electronically or optically reconfigured to alter its wavefront control properties.

References

1. Semantic regularization of electromagnetic inverse problems. Nature Communications (2024).
2. Transmission-Type 2-Bit Programmable Metasurface for Single-Sensor and Single-Frequency Microwave Imaging. Scientific Reports (2016).
3. Large Metasurface Aperture for Millimeter Wave Computational Imaging at the Human-Scale. Scientific Reports (2017).
4. Semantic–Electromagnetic Inversion With Pretrained Multimodal Generative Model. Advanced Science (2024).
5. Integrated convolutional neural networks for joint super-resolution and classification of radar images. Pattern Recognition (2024).
6. Learned Integrated Sensing Pipeline: Reconfigurable Metasurface Transceivers as Trainable Physical Layer in an Artificial Neural Network. Advanced Science (2019).