Scientific Reports

Super-resolved quantum ghost imaging

Chané Moodley, Andrew Forbes

Published Jun 20, 2022• 32 citations

Abstract

Quantum ghost imaging offers many advantages over classical imaging, including low photon fluxes and non-degenerate object and image wavelengths for imaging light sensitive structures, but suffers from slow image reconstruction speeds. Image reconstruction times depend on the resolution of the required image which scale quadratically with the image resolution. Here, we propose a super-resolved imaging approach based on neural networks where we reconstruct a low resolution image, which we denoise and super-resolve to a high resolution image. To test the approach, we implemented both a generative adversarial network as well as a super-resolving autoencoder in conjunction with an experimental quantum ghost imaging setup, demonstrating its efficacy across a range of object and imaging projective mask types. We achieved super-resolving enhancement of [Formula: see text] the measured resolution with a fidelity close to 90[Formula: see text] at an acquisition time of N[Formula: see text] measurements, required for a complete N [Formula: see text] N pixel image solution. This significant resolution enhancement is a step closer to a common ghost imaging goal, to reconstruct images with the highest resolution and the shortest possible acquisition time.

DOI: 10.1038/s41598-022-14648-2

Related papers

Quantum imaging gets super smart

Chané Moodley, Andrew Forbes

Quantum ghost imaging is an alternative imaging technique which utilises pairs of entangled photons to reconstruct an image. Due to the scanning nature of spatially resolving detectors and the inherent low light levels of quantum experiments, imaging speeds are inefficient and scale quadratically with the required resolution. We leveraged artificial intelligence capabilities to achieve early object recognition and to super-resolve the reconstructed image. We achieved a 5x reduction in image acquisition times and super-resolved the images to a resolution 4x greater than the measured resolution. Leading to efficient image acquisition times without losing fine details of the image.

Quantum Communications and Quantum ImagingOct 4, 20220 citations

Quantum ghost imaging with a modal basis

Fazilah Nothlawala, Chané Moodley, Neelan Gounden +2 more

Quantum ghost imaging harnesses entangled photon pairs to form images using light that never directly interacts with the object. Conventional approaches rely on pixelated detectors or spatial masks, where the resolution is fundamentally limited by the pixel size and thus bounded by the Rayleigh criterion. In this work, we present a modal approach that replaces pixel-based detection with measurements in a basis of structured modes. By reconstructing images through tailored spatial modes, we decouple resolution from the physical limitations of detectors and instead tie it to the system’s optical precision and computational control. We demonstrate that even non-orthogonal modal sets can achieve high fidelity image reconstruction by exploiting modal sparsity, significantly reducing the number of required measurements. Using phase-only approximations of Hermite-Gaussian modes, we avoid the loss associated with complex amplitude modulation while retaining high reconstruction fidelity. The result is a high-resolution, high-efficiency quantum ghost imaging technique powered by structured light, opening new possibilities for photon-efficient imaging in low-light and quantum-limited regimes, such as biological microscopy and remote sensing.

Optical and Quantum Sensing and Precision Metrology IIMar 5, 20260 citations

Time-efficient non-degenerate ghost imaging powered by deep learning

Chané Moodley, Bereneice Sephton, Valeria Rodríguez-Fajardo +1 more

Quantum ghost imaging utilizes entangled photon pairs to enable an alternative image acquisition method. Information from either one of the photons does not allow for image reconstruction, however the image can be reconstructed by utilising the correlations that exist between the photon pair. Interestingly, these photon pairs can be either degenerate or non-degenerate in nature. Non-degenerate ghost imaging offers the ability to image with wavelength bandwidths where spatially resolving detectors are impractical, ineffective or expensive. Due to the scanning nature of spatially resolving detectors and the inherent low light levels of quantum experiments, imaging speeds are rather unsatisfactory. To overcome this limitation, we propose a two-step deep learning approach to establish an optimal early stopping point, tested on a non-degenerate system. In step one, we enhance the reconstructed image after each measurement by a deep convolution auto-encoder, followed by step two where a classifier is used to recognise the image. We achieved a recognition confidence of 75% at 20% of the image reconstruction time, hence reducing the image reconstruction time 5-fold while maintaining the image information. This, therefore, leads to a faster, more efficient image acquisition technique. Although tested on a non-degenerate system, our procedure can be extended to many such systems that are of quantum nature. We believe that this two-step deep learning approach will prove valuable to the community who are focusing their efforts on time-efficient ghost imaging.

Quantum Communications and Quantum ImagingJul 30, 20211 citation

Smart Quantum Imaging

Ebrahim Karimi

The use of quantum properties of light has paved new avenues in the field of imaging. Quantum illumination and ghost-imaging protocols have shown exciting promise in addressing the issues inherent in classical imaging, such as background noise and a high level of required illumination. Moreover, quantum metrological schemes analyzing the modal content of light would address the resolution limits of direct imaging, promising enhancement beyond the Rayleigh limit. Augmenting these novel techniques with machine learning, particularly deep learning architectures such as Convolutional Neural Networks (CNNs), one strives to see significant improvements in image reconstruction and object identification in quantum imaging protocols. In my talk, I present recent progress and development on reconstructing the image of different objects with quantum sources, employing a high degree of spatial correlation between photon pairs and autoencoder deep learning architecture.

Photonics for Quantum (PfQ)Jul 15, 20210 citations

Machine learning assisted quantum super-resolution microscopy

Zhaxylyk A. Kudyshev, Demid Sychev, Zachariah O. Martin +7 more

One of the main characteristics of optical imaging systems is spatial resolution, which is restricted by the diffraction limit to approximately half the wavelength of the incident light. Along with the recently developed classical super-resolution techniques, which aim at breaking the diffraction limit in classical systems, there is a class of quantum super-resolution techniques which leverage the non-classical nature of the optical signals radiated by quantum emitters, the so-called antibunching super-resolution microscopy. This approach can ensure a factor of [Formula: see text] improvement in the spatial resolution by measuring the n -th order autocorrelation function. The main bottleneck of the antibunching super-resolution microscopy is the time-consuming acquisition of multi-photon event histograms. We present a machine learning-assisted approach for the realization of rapid antibunching super-resolution imaging and demonstrate 12 times speed-up compared to conventional, fitting-based autocorrelation measurements. The developed framework paves the way to the practical realization of scalable quantum super-resolution imaging devices that can be compatible with various types of quantum emitters.

Nature CommunicationsAug 10, 202324 citations

Does quantum ghost imaging need a camera to image an object?

Chané Moodley, Andrew Forbes

Pairs of entangled photos are used to reconstruct an image in the application area known as quantum ghost imaging. It is the correlation between the photon pair that allows for the reconstruction of the image, as opposed to single photon detection. The entangled photons are spatially separated into two independent paths, one to illuminate the object and the other which is collected by a spatially resolving detector. Initially, ghost imaging experiments accomplished spatially resolving detectors by moving a single-pixel detector throughout a transverse scanning area. Advancements consisted of using ultra-sensitive cameras to avoid a system consisting of physically moving detectors. Ultra-sensitive cameras are, however, expensive and have limited spectral sensitivity. Here we demonstrate an alternative by utilising a spatial light modulator and a bucket detector to spatially resolve what is detected. Importantly, the masks displayed on the spatial light modulator must constitute a complete basis to acquire a completely reconstructed image. Historically, imaging speeds have been slow and inefficient due to the quadratic increase in the scanning capability for spatially resolved detectors and the low light levels associated with quantum experiments. Here we additionally utilise machine and deep learning algorithms to improve both image reconstruction time and resolution. We demonstrate this with a non-degenerate ghost imaging setup where the physical parameters such as the mask type and resolution are varied and controlled on a spatial light modulator. Thereby answering the question: can we image an object without using a camera?

Sep 30, 20220 citations