This project aims to develop novel AI computational imaging methods to enable quantum-level sensing for high-quality vision in challenging conditions.

Single-photon detectors provide exceptional sensitivity and precise timing, making them ideal for 3D LiDAR and low-light imaging. They are uniquely suited for scenarios where conventional cameras struggle, such as underwater environments [1] or scenes obscured by fog, haze, or scattering media. Although recent AI methods have greatly improved classical imaging, they remain poorly adapted to the sparsity, noise, and physics of single-photon measurements.

This project will develop tailored AI solutions that combine statistical modelling with modern machine learning to produce high-resolution, interpretable, and efficient reconstructions from multidimensional single-photon data [2,3]. A particular emphasis will be placed on enabling reliable imaging in extreme environments (such as underwater, through fog, or in other visually degraded conditions) where scattering severely limits conventional vision systems. The student will design algorithms capable of operating in real time, fusing information from passive and active sensors, and leveraging the unique timing precision of single-photon detectors. The goal is to unlock the use of quantum-level sensing for critical applications such as underwater inspection, and autonomous navigation in poor visibility. More precisely, the objectives are:

  1. Develop computational imaging AI algorithmsthat deliver high-quality, interpretable, and fast reconstructions from sparse single-photon data using principled, physics-aware AI.
  2. Enable robust imaging in extreme environmentsby exploiting single-photon detection for underwater vision, imaging through fog or obscurants, and other degraded-visibility scenarios.

Expected Outcomes

The project will produce a new generation of AI-driven single-photon imaging methods that combine robustness, speed, and interpretability. These advances will significantly expand the practical capabilities of quantum sensing systems for inspection, navigation, and imaging in harsh environments.

Software Needs and Skills:  Statistical signal and Image processing, Bayesian methods, deep learning, optimization.  Python, Matlab.