Master intership + PhD at the Langevin Institute

Invariant Properties in Multimode Fibers for Imaging Applications

We are recruiting a master student with the possibility to continue during a Ph.D (funded) to work on the study of light propagation in multimode fibers using wavefront shaping and numerical reconstruction algorithms (phase retrieval, deep learning). Join un in Paris!

Keywords: waveftont shaping, mutlimode fibers, mesoscopic physics, phase retrieval, deep learning

See our recent publication: 

• [Learning and avoiding disorder in multimode fibers, Phys. Rev. X, 2021]
• [Tailoring the Rotational Memory Effect in Multimode Fibers, arxiv, 2023]

TL;DR:
We will play with deep learning frameworks to develop new approaches for calibration-less imaging through multimode fibers based on the study of invariant properties in multimode fibers.

Contact: Sébastien Popoff - sebastien.popoff(at)espci.fr

More information here.

Here is a small experimental chatbot designed to answer questions about wavefront shaping.

The model is trained on the full text transcript of wavefrontshaping.net, as well as a collection of hundreds of transcripts of wavefront shaping-related articles and abstracts.

Learning and Avoiding Disorder in Multimode Fibers

Sébastien M. Popoff
July 2021

In this work, we demonstrate the existence of a set of spatial channels in multimode fibers that are robust to strong local perturbations. We show that, even for a high level of disorder, light propagation can be characterized by just a few key properties.

Related article: doi.org/10.1103/PhysRevX.11.021060

Image reconstruction through unknown random configurations of multimode fibers using deep learning

[S. Resisi et al, arxiv, 2011.05144 (2020)]

Multimode fibers tend to scramble input images due to intermodal dispersion and random mode coupling. While this scrambling effect can be learned, for instance by measuring the transmission matrix of the fiber, slight changes of the geometrical conformation of the fiber modify its response, making the calibration obsolete. In the present paper, the authors use deep-learning using data sets acquired over a wide range of deformations to reconstruct images sent through unknown configurations of multimode fibers. This is possible thanks to the presence of invariant properties that the numerical model learns.