Machine learning assisted quantum super-resolution microscopy

Conventional optical imaging systems are limited by the diffraction limit, restricting spatial resolution to approximately half the wavelength of incident light. While classical super-resolution techniques aim to overcome this, quantum super-resolution microscopy leverages the non-classical nature of optical signals from quantum emitters (antibunching super-resolution microscopy). This approach improves spatial resolution by a factor of √*n* by measuring the *n*-th order autocorrelation function. However, acquiring multi-photon event histograms is time-consuming. This paper presents a machine learning-assisted approach for rapid antibunching super-resolution imaging, achieving a 12-times speedup compared to conventional methods. This framework facilitates the development of scalable quantum super-resolution imaging devices compatible with various quantum emitters.

Zhaxylyk A. Kudyshev, Demid Sychev, Zachariah Martin, Omer Yesilyurt, Simeon I. Bogdanov, Xiaohui Xu, Pei-Gang Chen, Alexander V. Kildishev, Alexandra Boltasseva, Vladimir M. Shalaev