Robust High-Resolution Imaging Through Bend-Resilient Multimode Fibres

Recent advancements in sensing and imaging technologies have driven innovation across various fields. However, the forward models governing these systems are often non-stationary, as sensor responses can fluctuate due to both internal and external factors. For instance, ultra-thin multimode fibres (MMFs) used in high-resolution imaging experience mode coupling, which alters unpredictably with each bend, distorting the transmitted signal. These challenges present a fundamental inverse problem: reconstructing accurate signals or images from inherently variable and noisy sensor outputs. Existing reconstruction methods often rely on assumptions of stable forward models, which limits their robustness under real-world, dynamic conditions. In this presentation, I will discuss my proposed algorithms for real-time imaging using MMFs, designed to maintain robustness against bending and other sources of variability.