arXiv · 2611.16094
Optical symmetry breaking for single-image field reconstruction
Abstract
Light carries information in both amplitude and phase, although cameras record only intensity. Reconstructing the optical field is at the heart of wavefront sensing, microscopy and coherent communications. Existing methods often rely on multiple measurements or iterative phase retrieval. In fields described by known spatial modes, intensity is determined by pairwise products of their complex coefficients. Yet optical symmetries can make these products indistinguishable, leaving different fields with nearly identical images. Here, we show that engineering the optical measurement to break these symmetries enables direct reconstruction of the complex field from a single intensity image. We optimise a phase mask followed by propagation to separate pairwise modal signatures while retaining light on the detector. A calibrated linear inverse and rank-one projection then recover modal amplitudes and relative phases without iterative phase retrieval. In simulations, optimised masks satisfy conditioning and light-capture requirements unmet by random masks. Experiments reconstruct 42-mode fibre fields using either a spatial light modulator or a passive diffuser, with median complex-field errors of 9.1% and 4.6%, respectively, against holographic references. This extends an earlier five-mode experimental demonstration of fast intensity-only modal decomposition. Per-frame computation takes 2.1 ms, and batched throughput exceeds 4,000 frames per second. These results establish measurement-space symmetry breaking as a route to rapid single-image optical field reconstruction.
Explore related subjects
Keep this discovery
Explore connections, maps & timelines
Egor Manuylovich. 2026-11-11. Optical symmetry breaking for single-image field reconstruction. https://searcharxiv.com/abs/2611.16094
Citations retain the original authors and DOI and include this site’s record URL. If this discovery service contributed to your research, you can also cite Searcharxiv by Indiciae: BibTeX · RIS.