Source-linked AI summary
Image Transmission Through an Opaque Material
S. M. Popoff, G. Lerosey, M. Fink, A. C. Boccara, S. Gigan
TL;DR
Multiple scattering makes image reconstruction through opaque media difficult to describe with classical means, despite deterministic propagation. The paper measures and exploits the transmission matrix to coherently recover arbitrary images, achieving more than 85% fidelity at γ = 11.
Problem
Real multiple-scattering propagation is too complex for classical description, leaving reconstruction of arbitrary images through opaque media as an inverse problem requiring transmission-matrix measurement.
Method
The approach measures the optical transmission matrix and uses its singular modes and noise-matched weighting to reconstruct transmitted images.
Results
More than 85% fidelity was achieved for the largest γ = 11 without averaging, while optimal reconstruction reached 93.6% and 94.5% correlation in two experimental conditions.
Takeaways & Limitations
The transmission matrix provides a general framework for rapid, accurate coherent imaging through strongly scattering linear complex media.
Takeaways & Limitations
Direct matrix inversion is highly unstable in noise, and singular values below the noise level can make reconstructed images unrelated to the input.
Abstract
from arXiv · showhide
Optical imaging relies on the ability to illuminate an object, collect and analyze the light it scatters or transmits. Propagation through complex media such as biological tissues was so far believed to degrade the attainable depth as well as the resolution for imaging because of multiple scattering. This is why such media are usually considered opaque. Very recently, we have proven that it is possible to measure the complex mesoscopic optical transmission channels that allows light to traverse through such an opaque medium. Here we show that we can optimally exploit those channels to coherently transmit and recover with a high fidelity an arbitrary image, independently of the complexity of the propagation.
Methods · Imaging Setup
The imaging setup used a spatially modulated 532 nm laser focused through an opaque ZnO scattering medium. The medium was an 80 ± 25 µm deposit with a measured transport mean free path of 6±2 microns.
- Imaging Setup: A 532 nm laser source provided the incident light for the experiment.The source was identified as a Laser Quantum Torus.
- Imaging Setup: The incident beam was expanded before spatial modulation and focusing.These beam-conditioning steps preceded propagation through the scattering medium.
- Imaging Setup: A Spatial Light Modulator spatially modulated the expanded beam.The modulator was a Holoeye LC-R 2500.
- Imaging Setup: The modulated light was focused on an opaque, strongly scattering medium.The medium consisted of a ZnO deposit on a standard microscope glass slide.
- Imaging Setup: 80 ± 25 µm was the thickness of the ZnO deposit used as the scattering medium.The deposit was specified as Sigma-Aldrich 96479.
- Imaging Setup: 6±2 microns was the measured transport mean free path of the ZnO medium.The measurement characterized the scattering deposit used in the optical setup.
Generation of the amplitude object
The method generates a virtual amplitude object by subtracting two phase objects, avoiding the need to directly control the incident beam’s amplitude and phase. A second phase mask shifts one pixel relative to the first, enabling amplitude estimation from the resulting output speckle.
- Generation of the amplitude object: A virtual amplitude object is generated by subtracting two phase objects because the incident beam’s amplitude and phase cannot be simply controlled.The virtual object is denoted Eobj with amplitude values sobj_m ∈ [0, 1].
- Generation of the amplitude object: The virtual-object method is more flexible than placing a real amplitude object in the spatial-light-modulator plane.
- Generation of the amplitude object: From any phase mask E(1)_phase, a second mask E(2)_phase is generated by shifting the mth pixel’s phase by sobj_m.