Research · Sep 2025 to now

Phase recovery with a phone as the light source

Fourier Ptychographic Microscopy (FPM) recovers phase and improves resolution by imaging a sample under many illumination angles. I replaced its programmable LED array with an ordinary smartphone screen.

Computational and Wave Imaging Lab, Chandra Family Department of Electrical and Computer Engineering, The University of Texas at Austin. Research assistant to Prof. Shwetadip Chowdhury.

The poster

Full-size image · PDF

Why a phone

Standard FPM uses a programmable LED array, with dozens to hundreds of individually addressable elements, as the angular-diversity source. That hardware is expensive and hard to access outside a specialized optics lab.

I replaced the LED array with an ordinary smartphone screen. Sweeping a bright point across the display provides the same angularly diverse illumination at near-zero hardware cost. Each screen position acts as a “virtual LED.” I built the acquisition sequencing and the full EPRY-based reconstruction pipeline around this substitution.

This allows the method to be used outside the lab, without complex and expensive cameras and LEDs.

Planned experimental setup

Diagram of the planned setup: a phone screen, the sample, lens L1, the Fourier plane with an iris, lens L2 and a camera, arranged as a 4f relay — Phone-illuminated FPM acquisition. The phone screen replaces the LED array. Standoff distance and screen pitch set the achievable illumination NA.

A 4f relay with two 50 mm lenses (1× magnification) and an iris at the shared Fourier plane, with no objective.
The phone sits behind the sample and displays a moving bright point, stepping through a grid of N = 200 positions, each giving a different illumination angle.
A script drives the phone and triggers the camera, so each frame maps to one known illumination angle.
A single color channel (blue) from the phone gives a well-defined illumination wavelength.
The phone-to-sample distance is fixed with a sample stage to set the illumination NA.
Each screen position's illumination angle is calibrated against a known target before acquisition.

Reconstruction pipeline

Display a bright point at each of N calibrated screen positions, capturing one low-resolution intensity frame per position.
Take the FFT of each capture and place it in the correct location in the full Fourier spectrum, based on the illumination angle for that position.
Crop the matching patch from the current image estimate, replace its magnitude with the measured magnitude (keeping the estimate's phase), then transform back and update that region of the estimate.
Repeat across all N images until the estimate stops changing significantly, then assemble the final high-resolution amplitude and phase reconstruction.

Plot of overlapping circular patches tiled across the Fourier plane, with a large red circle labeled 1.4 NA and a small blue circle labeled 0.3 NA — Fourier synthesis. Overlapping spectrum patches, one per illumination angle. Red circle: 1.4 NA. Blue circle: 0.3 NA.

Results

Ten low-resolution grayscale microscope captures of cells, labeled LED 0 through LED 9 — Raw low-resolution intensity captures, one per phone-screen illumination position. This is the direct input to the reconstruction.

Reconstructed amplitude image, mostly flat gray, beside the reconstructed phase image in radians showing detailed cell structure — Reconstructed amplitude and reconstructed phase (radians).

What I observed

Phone-screen brightness is lower than a dedicated LED, which makes keeping the signal-to-noise ratio usable a challenge.
Screen pixel pitch and standoff distance jointly cap the achievable illumination NA. I mapped this for my setup.
Position-to-angle calibration is critical. Small errors visibly blur the EPRY-recovered pupil.

A lab bench with the optical setup, a laptop and an orange instrument in front of a whiteboard covered in equations — The bench in the lab.

Conclusions

A smartphone screen is a fully functional, near-zero-cost, programmable angular-diversity source for FPM. It reproduces standard LED-array acquisition geometry.
This substitution meaningfully lowers the hardware barrier to quantitative phase imaging.

References

G. Zheng, R. Horstmeyer, C. Yang. Wide-field, high-resolution Fourier ptychographic microscopy. Nat. Photonics 7, 739–745 (2013).
X. Ou, G. Zheng, C. Yang. Embedded pupil function recovery for Fourier ptychographic microscopy. Opt. Express 22, 4960–4972 (2014).
L. Tian, X. Li, K. Ramchandran, L. Waller. Multiplexed coded illumination for Fourier ptychography. Biomed. Opt. Express 5, 2376–2389 (2014).

← Back to all projects