Accurate and fast determination of the topological charge of Laguerre–Gauss entangled photons with convolutional neural networks
Year: 2026
Type: article
Abstract: In this manuscript, we report the implementation of a highly efficient and automated system capable of determining the topological charge of photon pairs, entangled in their orbital angular momentum (OAM) degree of freedom and projected in the Laguerre-Gauss (LG) basis. The system relies on a convolutional neural network (CNN), trained to classify the transverse intensity distribution of heralded LG photons, transformed by a tilted bi-convex lens into Hermite-Gauss (HG) like patterns of intensity lobes. Our results show that the CNN is capable of determining the topological charge (magnitude and sign) of the idler photon in a LG mode, with values up to ℓ ± 10 , generating a diagonal confusion matrix with a classification speed of 3263 images/s, a mean square error of 0.0712, and accuracy about 97.9% in the testing stage. We believe our results represent a step forward for the generalized integration of CNN’s in future quantum imaging and communication protocols.
Source: Results in Optics
Language: English
Subfield: Atomic and Molecular Physics, and Optics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: gold
Estimated APC paid: $1,470
OpenAlex ID: W7204240231
