Scaling Optical Throughput via Distributed Aperture Arrays

24 concepts16 papers1 proposal

In classical optics, a stubborn trade-off dictates that an imaging system can have high spatial resolution or a wide field of view (FOV), but rarely both. If you want to resolve sub-micron structures, you need a high numerical aperture (NA) lens, which inherently restricts your view to a tiny window. Conversely, expanding the FOV to capture large biological tissues or expansive skies usually means sacrificing the ability to see fine details.

To break this bottleneck, a powerful paradigm has emerged: distributed aperture and lensless computational imaging. Instead of forcing a single, massive physical lens to collect and focus all light rays perfectly, these systems distribute the optical workload across arrays of smaller lenses, sensors, or phase masks. By shifting the burden of image formation from complex glass manufacturing to digital algorithms, this approach achieves gigapixel-scale throughput across centimeters or kilometers.

The Tyranny of the Single Large Lens

Why can't we just build a giant lens that does everything? As a lens increases in size and field of view, optical aberrations—such as chromatic distortion and off-axis blurring—scale exponentially. When a single wavefront is stretched across a large aperture, off-axis light rays strike the lens edges at steep angles, causing severe phase errors that drop spatial information below the sensor's noise floor.

To maintain a stable Modulation Transfer Function (MTF) across a wide field, traditional monolithic systems must resort to extreme engineering. For instance, Taro Ichimura et al. initially paired a low-power telecentric lens with a high-resolution CMOS sensor in Exploring rare cellular activity in more than one million cells by a transscale scope to image a million cells simultaneously. However, the authors later found that achieving true volumetric sub-cellular resolution across a centimeter-wide tissue required a custom-designed giant lens containing 19 separate optical elements, as detailed in Volumetric trans-scale imaging of massive quantity of heterogeneous cell populations in centimeter-wide tissue and embryo. Similarly, Adam K. Glaser et al. in Expansion-assisted selective plane illumination microscopy for nanoscale imaging of centimeter-scale tissues had to repurpose massive, high-etendue industrial metrology lenses designed for circuit board inspection just to achieve nanoscale light-sheet imaging of intact specimens without physical sectioning.

These monolithic approaches are physically bulky, prohibitively expensive, and optically constrained.

Mitigating Aberrations via Tiled Micro-Apertures

The core intuition behind distributed aperture imaging is elegant: divide and conquer. If you replace a single large aperture with an array of smaller optical elements, each sub-aperture only has to handle a narrow, highly localized field of view. Within this narrow window, light rays remain close to the optical axis, meaning aberrations are minimal and the MTF remains high and stable.

Xiong Dun et al. demonstrated this principle in Neural array meta-imaging using a metalens array paired with deep learning. By breaking down a large aperture into smaller sub-apertures, they prevented wavefront aberrations from destroying high-frequency spatial information. Because the optical information is preserved above the sensor's noise floor, a neural network can successfully reconstruct a crisp, full-color image, shrinking the optical track length by a factor of 13 compared to bulky commercial lenses.

This modular strategy scales seamlessly. In microscopy, Eric E Thomson et al. developed the Multi-Camera Array Microscope (MCAM) in Gigapixel imaging with a novel multi-camera array microscope, which tiles dozens of inexpensive CMOS sensors—each equipped with its own micro-lens—into a large array. Because each camera independently captures a local patch without off-axis distortion, the total system resolution scales linearly with the number of sensors, enabling parallel acquisition of gigapixel-scale images and videos of freely moving organisms.

The same logic applies across vastly different domains:

Coherent Synthesis and Remote Focusing

When sub-apertures capture overlapping fields coherently, computation can do more than just stitch images—it can synthesize a larger virtual aperture. Jason Holloway demonstrated in Toward Long-Distance Subdiffraction Imaging Using Coherent Camera Arrays that high-resolution, long-distance imaging can be achieved more cost-effectively through algorithmic synthesis than through traditional high-precision lens manufacturing. By distributing the aperture across multiple independent optical paths, the system bypasses traditional physical constraints.

In 3D biological imaging, managing these distributed optical paths requires precise control over the focal plane. Single-objective light-sheet systems traditionally struggle to maintain high resolution across large volumes due to relay constraints. To solve this, Bin Yang et al. introduced a custom tertiary objective lens into a remote focusing optical train in DaXi—high-resolution, large imaging volume and multi-view single-objective light-sheet microscopy. This design maintains a high numerical aperture across millimeter-scale volumes, eliminating the traditional trade-off between a large field of view and high numerical aperture, enabling the imaging of multiple organisms simultaneously without loss of detail.

The Ultimate Limit: Lensless Computational Imaging

If distributing the aperture across small lenses mitigates aberrations, what happens if we remove the lenses entirely? This is the premise of lensless microscopy. Without a lens, light scattered from a sample propagates directly onto a sensor array, creating a wide-field interference pattern. Because there is no physical lens to limit the field of view, high-throughput screening of large fluid volumes becomes possible. Yibo Zhang et al. leveraged this in Motility-based label-free detection of parasites in bodily fluids using holographic speckle analysis and deep learning to detect rare, motile parasites in low-concentration samples by capturing wide-field holographic speckle patterns.

To recover sharp spatial details from these raw diffraction patterns, researchers introduce structured phase masks close to the sensor. For example, Jesse K. Adams et al. developed the "Bio-FlatScope" in In vivo lensless microscopy via a phase mask generating diffraction patterns with high-contrast contours, which utilizes a phase mask designed to generate high-contrast contours. These contours ensure that even low-contrast biological tissues produce highly modulated diffraction patterns that can be computationally inverted to achieve a lateral resolution under 9 µm and an axial depth range of 500 µm in vivo.

[Sample] ---> [Phase Mask / Diffraction] ---> [Sensor Array] ---> [GMM / Super-Resolution] ---> [High-Res Image]

To prevent computational reconstruction from becoming a temporal bottleneck, the architecture can be parallelized. Xin Yuan et al. proposed a modular sensor array architecture in Block-wise Lensless Compressive Camera, where each sensor captures its own local patch independently and simultaneously. This allows the total resolution to scale with the number of sensors without increasing capture time. Similarly, You Zhou et al. combined rapid burst acquisition with pixel super-resolution and region-wise parallel processing in Large-scale lensless microscopy with fast acquisition and region-wise focusing, achieving a sub-micron resolution across a field of view exceeding 40 mm² in just 0.5 seconds.

Conclusion

By shifting the core mechanism of image formation from physical refraction to computational synthesis, distributed aperture and lensless systems bypass the geometric constraints of classical lenses. Whether by tiling micro-cameras to capture gigapixel biological dynamics or using phase masks to eliminate lenses entirely, these technologies demonstrate that the future of high-throughput imaging lies not in heavier glass, but in smarter algorithms and distributed arrays.

Key Foundational Work

MULTI-CAMERA LAPAROSCOPIC IMAGING WITH TUNABLE FOCUSING CAPABILITY (2014) — contributed 2 concept elements to this direction.
Design and characterization of thin multiple aperture infrared cameras (2009) — contributed 2 concept elements to this direction.
History & citationVersion 1 · 2026-10-04

Current · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_distributed_aperture_imaging_16d7dfaa,
  title = {distributed aperture imaging},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/distributed-aperture-imaging/rdr_16d7dfaaedfd494caf000c554e3036d6},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}

This exact version (version 1, frozen text):

@misc{lacuna_distributed_aperture_imaging_16d7dfaa_v1,
  title = {distributed aperture imaging},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/distributed-aperture-imaging/rdr_16d7dfaaedfd494caf000c554e3036d6/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}
About

These directions overlap with this one through shared papers.

See something wrong? Report a mistake