Abstract Object identification is central to modern sensing, widely employed in autonomous systems and industrial inspection. However, conventional approaches rely on image formation followed by computational inference, requiring many photons to suppress shot noise and introducing readout and computational latencies. These limitations arise because direct position-basis measurement is not optimized for discriminating object-encoded states. Here, we demonstrate single-photon object identification without imaging by implementing quantum state discrimination in a high-dimensional spatial Hilbert space. By applying a task-optimized unitary transformation prior to wavefunction collapse, we transform spatially overlapping object-encoded wavefields into spatially separated states. Thus, the measurement basis is engineered so that the collapse of the single-photon state directly specifies the object label. Numerically, our method achieves 90.2% identification accuracy for nine objects, surpassing the 31.9% classical limit obtained from maximum-likelihood single-photon intensity imaging. Experimentally, the framework achieves 71.0% conditional identification accuracy under attenuated coherent illumination with camera detection, and is further demonstrated with heralded single photons using only 36 detection channels. Our framework establishes a quantum-measurement approach to information-efficient sensing that connects quantum state discrimination with computational optics, providing a route toward low-photon and ultrafast object identification.
Abstract Object identification is central to modern sensing, widely employed in autonomous systems and industrial inspection. However, conventional approaches rely on image formation followed by computational inference, requiring many photons to suppress shot noise and introducing readout and computational latencies. These limitations arise because direct position-basis measurement is not optimized for discriminating object-encoded states. Here, we demonstrate single-photon object identification without imaging by implementing quantum state discrimination in a high-dimensional spatial Hilbert space. By applying a task-optimized unitary transformation prior to wavefunction collapse, we transform spatially overlapping object-encoded wavefields into spatially separated states. Thus, the measurement basis is engineered so that the collapse of the single-photon state directly specifies the object label. Numerically, our method achieves 90.2% identification accuracy for nine objects, surpassing the 31.9% classical limit obtained from maximum-likelihood single-photon intensity imaging. Experimentally, the framework achieves 71.0% conditional identification accuracy under attenuated coherent illumination with camera detection, and is further demonstrated with heralded single photons using only 36 detection channels. Our framework establishes a quantum-measurement approach to information-efficient sensing that connects quantum state discrimination with computational optics, providing a route toward low-photon and ultrafast object identification.
Image classification is a core task of intelligent sensing, conventionally follows a sequential imaging then processing pipeline. However, redundant high-dimensional image reconstruction is inherently inefficient, especially in photon limited scenarios. Here we report a photon level image classification method using quantum compressed sensing, which reformulates the classification task as a sparse signal measurement problem directly oriented toward class labels. By exploiting the parallelism of photonic quantum superposition states, a single photon can be encoded the complete spatial information of a high-dimensional image. Through a diffractive deep neural network, we physically construct a dedicated measurement basis aligned with the class space, enabling signal-dependent adaptive compressive measurement. Ideally, our method can extract class information via a single quantum projective measurement, reducing the required number of measurements from the logarithmic scaling O(Klog(N/K)) of classical compressed sensing to the constant-order information-theoretic limit M = K = 1. Experimental results show that a classification accuracy of 69.0% can be achieved by using a single-photon detection event as the decision criterion, while it increases to 95.0% with four-photon detection events. This work demonstrates image classification at the energy efficiency limit and introduces a measurement as decision framework. It provides a foundation for intelligent sensing systems that operate under extreme photon budgets and harsh environments.
The rapid growth of machine learning is increasingly constrained by the energy and bandwidth limits of classical hardware. Optical and quantum technologies offer an alternative route, enabling high-dimensional, parallel information processing directly in the physical layer, particularly suited for imaging tasks. In this context, quantum photonic platforms provide both a natural mechanism for computing inner products and a promising path to energy-efficient inference in photon-limited regimes. Here, we experimentally demonstrate a camera-free quantum-optical images classifier that performs inference directly at the measurement layer using Hong-Ou-Mandel (HOM) interference of spatially programmable single photons. Two-photon coincidences directly report the overlap between an input image mode and a learned template, replacing pixel-resolved acquisition with a single global measurement. We realize both a single-perceptron quantum optical neuron and a two-neuron shallow network, achieving high accuracy on benchmark datasets with strong robustness to experimental noise and minimal hardware complexity. With a fixed measurement budget, performance remains insensitive to input resolution, demonstrating intrinsic robustness to the number of pixels, which would be impossible in a classical framework. This approach paves the way toward neuromorphic quantum photonic processors capable of extracting task-relevant information directly from HOM interference, with promising applications in remote object recognition, low-signal sensing, and photon-starved biological microscopy.
Photonic quantum technologies utilize various degrees of freedom (DOFs) of light, such as polarization, frequency, and spatial modes, to encode quantum information. In the effort of further improving channel capacity of quantum communication, and for increasing the complexity of available quantum operations, high-dimensional and hyperentangled states are now gaining interest. However, efficiently measuring these high dimensional states is challenging due to the large number of measurements required for reconstructing the full density matrix via quantum state tomography (QST), and the fact that each measurement requires some modification in the experimental setup. Here, we propose a framework for reconstructing the density matrix of a single-photon hyperentangled across multiple DOFs using a single intensity-measurement obtainable from traditional cameras, and discuss extensions for multiphoton hyperentangled states. Our method hinges on the spatial DOF of the photon and uses it to encode the quantum information from the other DOFs. We numerically demonstrate this method for single-photon OAM-spin and OAM-frequency entangled states using an ideal coupler and a multimode fiber, to perform the information mixing and transfer the encoding to spatial information, where it is detected using a simple camera. This technique simplifies the experimental setup and reduces acquisition time compared to traditional QST-based methods. Moreover, it allows the recovery of DOFs that conventional cameras cannot detect, such as polarization, thus eliminating the need for projection measurements.
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.
Abstract Single-pixel imaging is a method for obtaining an image without multi-pixel sensors. In this method, an object of interest is illuminated by light that has specifically designed spatio-temporal shape, while a single-pixel detector measures the total amount of reflected or transmitted light. The key prerequisite of this method is the ability to create a given spatio-temporal light shape, which becomes especially challenging at the level of single photons. The current photon shaping procedures cannot be directly applied to single-pixel imaging due to the limited generation rate or the need for post-processing. In this work, we propose a highly efficient method for generating arbitrary single-photon shapes without post-selection. We exploit the heralded single-photon shaping approach, and study the effect of the modal properties of a single-photon detection on the characteristics of the heralded photon. We find that the use of multi-mode detection allows to increase the heralding probability of shaped photons proportionally to the number of detected modes while keeping the shape intact, which makes it highly desirable for single-pixel quantum imaging.
Abstract Measurements approaching the ultimate quantum limits of sensitivity are central in quantum information processing, quantum metrology, and communication. Quantum measurements to discriminate multiple states at the single-photon level are essential for optimizing information transfer in low-power optical communications and quantum communications, and can enhance the capabilities of many quantum information protocols. Here, we theoretically investigate and experimentally demonstrate the discrimination of multiple coherent states of light with sensitivities surpassing the quantum noise limit (QNL) at the single-photon level under realistic conditions of loss and noise based on strategies implementing globally-optimized adaptive measurements with single photon counting and displacement operations. These discrimination strategies can provide realistic advantages to enhance information transfer at low powers, and are compatible with photon number resolving detection, which provides robustness at high powers, thus allowing for surpassing the QNL at arbitrary input power levels under realistic conditions.