All-analog photoelectronic chip for high-speed vision tasks.

Photonic computing enables faster and more energy-efficient processing of vision data1-5. However, experimental superiority of deployable systems remains a challenge because of complicated optical nonlinearities, considerable power consumption of analog-to-digital converters (ADCs) for downstream digital processing and vulnerability to noises and system errors1,6-8. Here we propose an all-analog chip combining electronic and light computing (ACCEL). It has a systemic energy efficiency of 74.8 peta-operations per second per watt and a computing speed of 4.6 peta-operations per second (more than 99% implemented by optics), corresponding to more than three and one order of magnitude higher than state-of-the-art computing processors, respectively. After applying diffractive optical computing as an optical encoder for feature extraction, the light-induced photocurrents are directly used for further calculation in an integrated analog computing chip without the requirement of analog-to-digital converters, leading to a low computing latency of 72 ns for each frame. With joint optimizations of optoelectronic computing and adaptive training, ACCEL achieves competitive classification accuracies of 85.5%, 82.0% and 92.6%, respectively, for Fashion-MNIST, 3-class ImageNet classification and time-lapse video recognition task experimentally, while showing superior system robustness in low-light conditions (0.14 fJ µm-2 each frame). ACCEL can be used across a broad range of applications such as wearable devices, autonomous driving and industrial inspections.

Two-photon synthetic aperture microscopy for minimally invasive fast 3D imaging of native subcellular behaviors in deep tissue.

Holistic understanding of physio-pathological processes requires noninvasive 3D imaging in deep tissue across multiple spatial and temporal scales to link diverse transient subcellular behaviors with long-term physiogenesis. Despite broad applications of two-photon microscopy (TPM), there remains an inevitable tradeoff among spatiotemporal resolution, imaging volumes, and durations due to the point-scanning scheme, accumulated phototoxicity, and optical aberrations. Here, we harnessed the concept of synthetic aperture radar in TPM to achieve aberration-corrected 3D imaging of subcellular dynamics at a millisecond scale for over 100,000 large volumes in deep tissue, with three orders of magnitude reduction in photobleaching. With its advantages, we identified direct intercellular communications through migrasome generation following traumatic brain injury, visualized the formation process of germinal center in the mouse lymph node, and characterized heterogeneous cellular states in the mouse visual cortex, opening up a horizon for intravital imaging to understand the organizations and functions of biological systems at a holistic level.

Iterative tomography with digital adaptive optics permits hour-long intravital observation of 3D subcellular dynamics at millisecond scale.

Long-term subcellular intravital imaging in mammals is vital to study diverse intercellular behaviors and organelle functions during native physiological processes. However, optical heterogeneity, tissue opacity, and phototoxicity pose great challenges. Here, we propose a computational imaging framework, termed digital adaptive optics scanning light-field mutual iterative tomography (DAOSLIMIT), featuring high-speed, high-resolution 3D imaging, tiled wavefront correction, and low phototoxicity with a compact system. By tomographic imaging of the entire volume simultaneously, we obtained volumetric imaging across 225 × 225 × 16 µm3, with a resolution of up to 220 nm laterally and 400 nm axially, at the millisecond scale, over hundreds of thousands of time points. To establish the capabilities, we investigated large-scale cell migration and neural activities in different species and observed various subcellular dynamics in mammals during neutrophil migration and tumor cell circulation.

Sinusoidal Sampling Enhanced Compressive Camera for High Speed Imaging.

Compressive sensing technique allows capturing fast phenomena at a much higher frame rate than the camera sensor, by recovering a frame sequence from their encoded combination. However, most conventional compressive video sensing methods limit the achieved frame rate improvement to tenfold and only support low resolution recovery. Making use of the camera's redundant spatial resolution for further frame rate improve, here we report a novel compressive video acquisition technique termed Sinusoidal Sampling Enhanced Compressive Camera (S2EC2) to encode denser frames within a snapshot. Specifically, we decompose the dense frames into groups and apply combinational coding: random codes within each group for compressive acquisition; group specific sinusoidal codes to multiplex different groups onto the high resolution sensor. The sinusoidal codes designed for these groups would shift their frequency components by different offsets in the Fourier domain and staggered the dominant frequencies of the coded measurements of these groups. Correspondingly, the reconstruction successfully separate coded measurements of different groups and recovers frames within each group. Besides, we also solve the implementation problem of insufficient gray scale spatial light modulation speed, and build a prototype achieving 2000 fps reconstruction with a 15.6 fps camera (the actual compression ratio is 0.009). The extensive experiments validate the proposed approach.

Symmetry-breaking-induced dynamics in a nonlinear microresonator.

Optically induced symmetry-breaking plays a key role in nonlinear photonics. Recently, the experiment has successfully observed the Kerr-nonlinearity-induced chiral symmetry breaking in a single ultrahigh-Q whispering-gallery microresonator. Here, we show this symmetry-breaking can generate exotic dynamics between two counter-propagating modes. In particular, we predict two kinds of self-trappings, in which the corresponding relative phase oscillates around π or runs without bound although they have both the nonzero mean energy imbalance. Finally, we also clarify the impacts of the mode loss, finding a dynamical transition from self-trappings to an anharmonic oscillation. The presented scheme offers a new route to understanding the nonlinear dynamics and wave chaos in the microresonator.

Fourier-space Diffractive Deep Neural Network.

In this Letter we propose the Fourier-space diffractive deep neural network (F-D^{2}NN) for all-optical image processing that performs advanced computer vision tasks at the speed of light. The F-D^{2}NN is achieved by placing the extremely compact diffractive modulation layers at the Fourier plane or both Fourier and imaging planes of an optical system, where the optical nonlinearity is introduced from ferroelectric thin films. We demonstrated that F-D^{2}NN can be trained with deep learning algorithms for all-optical saliency detection and high-accuracy object classification.

Phase-space deconvolution for light field microscopy.

Light field microscopy, featuring with snapshot large-scale three-dimensional (3D) fluorescence imaging, has aroused great interests in various biological applications, especially for high-speed 3D calcium imaging. Traditional 3D deconvolution algorithms based on the beam propagation model facilitate high-resolution 3D reconstructions. However, such a high-precision model is not robust enough for the experimental data with different system errors such as optical aberrations and background fluorescence, which bring great periodic artifacts and reduce the image contrast. In order to solve this problem, here we propose a phase-space deconvolution method for light field microscopy, which fully exploits the smoothness prior in the phase-space domain. By modeling the imaging process in the phase-space domain, we convert the spatially-nonuniform point spread function (PSF) into a spatially-uniform one with a much smaller size. Experiments on various biological samples and resolution charts are demonstrated to verify the contrast enhancement with much fewer artifacts and 10-times less computational cost by our method without any hardware modifications required.

Snapshot quantitative phase microscopy with a printed film.

This paper proposes a low-cost snapshot quantitative phase imaging approach. The setup is simple and adds only a printed film to a conventional microscope. The phase of a sample is regarded as an additional aberration of the optical imaging system. And the image captured through a phase object is modeled as the distorted version of a projected pattern. An optimization algorithm is utilized to recover the phase information via distortion estimation. We demonstrate our method on various samples such as a micro-lens array, IMR90 cells and the dynamic evaporation process of a water drop, and our approach has a capability of real-time phase imaging for highly dynamic phenomenon using a traditional microscope.

GPU-based deep convolutional neural network for tomographic phase microscopy with ℓ1 fitting and regularization.

Tomographic phase microscopy (TPM) is a unique imaging modality to measure the three-dimensional refractive index distribution of transparent and semitransparent samples. However, the requirement of the dense sampling in a large range of incident angles restricts its temporal resolution and prevents its application in dynamic scenes. Here, we propose a graphics processing unit-based implementation of a deep convolutional neural network to improve the performance of phase tomography, especially with much fewer incident angles. As a loss function for the regularized TPM, the ℓ1-norm sparsity constraint is introduced for both data-fidelity term and gradient-domain regularizer in the multislice beam propagation model. We compare our method with several state-of-the-art algorithms and obtain at least 14 dB improvement in signal-to-noise ratio. Experimental results on HeLa cells are also shown with different levels of data reduction.

Quantum mixed phases of a two-dimensional polarized degenerate Fermi gas in an optical cavity.

The coupling of ultracold fermions to a high-finesse optical cavity can result in novel many-body phenomena, and has attracted significant interests at present. Here we consider a realization of the Fermi-Dicke model with controllable parameters, based on a two-dimensional polarized degenerate Fermi gas coupled to an optical cavity. We analytically investigate the ground-state properties of such system under the mean-field approximation. We find the system can exhibit a rich phase diagram depending on the fermion-photon coupling strength and the atomic resonant frequency. Contrasting to the bosonic counterpart, a first-order quantum phase transition between the superradiant phase and the normal phase featuring two Fermi surfaces can occur for the weak atomic resonant frequency, and there is a unique mixed phase where this normal phase and the superradiant phase coexist. The experimental detection of our results is also discussed.

Efficient single-pixel multispectral imaging via non-mechanical spatio-spectral modulation.

Combining spectral imaging with compressive sensing (CS) enables efficient data acquisition by fully utilizing the intrinsic redundancies in natural images. Current compressive multispectral imagers, which are mostly based on array sensors (e.g, CCD or CMOS), suffer from limited spectral range and relatively low photon efficiency. To address these issues, this paper reports a multispectral imaging scheme with a single-pixel detector. Inspired by the spatial resolution redundancy of current spatial light modulators (SLMs) relative to the target reconstruction, we design an all-optical spectral splitting device to spatially split the light emitted from the object into several counterparts with different spectrums. Separated spectral channels are spatially modulated simultaneously with individual codes by an SLM. This no-moving-part modulation ensures a stable and fast system, and the spatial multiplexing ensures an efficient acquisition. A proof-of-concept setup is built and validated for 8-channel multispectral imaging within 420~720 nm wavelength range on both macro and micro objects, showing a potential for efficient multispectral imager in macroscopic and biomedical applications.

Electric-field-induced interferometric resonance of a one-dimensional spin-orbit-coupled electron.

The efficient control of electron spins is of crucial importance for spintronics, quantum metrology, and quantum information processing. We theoretically formulate an electric mechanism to probe the electron spin dynamics, by focusing on a one-dimensional spin-orbit-coupled nanowire quantum dot. Owing to the existence of spin-orbit coupling and a pulsed electric field, different spin-orbit states are shown to interfere with each other, generating intriguing interference-resonant patterns. We also reveal that an in-plane magnetic field does not affect the interval of any neighboring resonant peaks, but contributes a weak shift of each peak, which is sensitive to the direction of the magnetic field. We find that this proposed external-field-controlled scheme should be regarded as a new type of quantum-dot-based interferometry. This interferometry has potential applications in precise measurements of relevant experimental parameters, such as the Rashba and Dresselhaus spin-orbit-coupling strengths, as well as the Landé factor.

Phase-factor-dependent symmetries and quantum phases in a three-level cavity QED system.

Unlike conventional two-level particles, three-level particles may support some unitary-invariant phase factors when they interact coherently with a single-mode quantized light field. To gain a better understanding of light-matter interaction, it is thus necessary to explore the phase-factor-dependent physics in such a system. In this report, we consider the collective interaction between degenerate V-type three-level particles and a single-mode quantized light field, whose different components are labeled by different phase factors. We mainly establish an important relation between the phase factors and the symmetry or symmetry-broken physics. Specifically, we find that the phase factors affect dramatically the system symmetry. When these symmetries are breaking separately, rich quantum phases emerge. Finally, we propose a possible scheme to experimentally probe the predicted physics of our model. Our work provides a way to explore phase-factor-induced nontrivial physics by introducing additional particle levels.

Multispectral imaging using a single bucket detector.

Existing multispectral imagers mostly use available array sensors to separately measure 2D data slices in a 3D spatial-spectral data cube. Thus they suffer from low photon efficiency, limited spectrum range and high cost. To address these issues, we propose to conduct multispectral imaging using a single bucket detector, to take full advantage of its high sensitivity, wide spectrum range, low cost, small size and light weight. Technically, utilizing the detector's fast response, a scene's 3D spatial-spectral information is multiplexed into a dense 1D measurement sequence and then demultiplexed computationally under the single pixel imaging scheme. A proof-of-concept setup is built to capture multispectral data of 64 pixels × 64 pixels × 10 wavelength bands ranging from 450 nm to 650 nm, with the acquisition time being 1 minute. The imaging scheme holds great potentials for various low light and airborne applications, and can be easily manufactured as production-volume portable multispectral imagers.

Photon Devil's staircase: photon long-range repulsive interaction in lattices of coupled resonators with Rydberg atoms.

The realization of strong coherent interactions between individual photons is a long-standing goal in science and engineering. In this report, based on recent experimental setups, we derive a strong photon long-range repulsive interaction, by controlling the van der Waals repulsive force between Cesium Rydberg atoms located inside different cavities in extended Jaynes-Cummings-Hubbard lattices. We also find novel quantum phases induced by this photon long-range repulsive interaction. For example, without photon hopping, a photon Devil's staircase, induced by the breaking of long-range translation symmetry, can emerge. If photon hopping occurs, we predict a photon-floating solid phase, due to the motion of particle- and hole-like defects. More importantly, for a large chemical potential in the resonant case, the photon hopping can be frozen even if the hopping term exists. We call this new phase the photon-frozen solid phase. In experiments, these predicted phases could be detected by measuring the number of polaritons via resonance fluorescence.

Polymer-grafted nanoparticles prepared by surface-initiated polymerization: the characterization of polymer chain conformation, grafting density and polydispersity correlated to the grafting surface curvature.

The polymer-grafted nanoparticles prepared by the surface-initiated polymerization induced from the spherical surface is studied by coarse-grained molecular dynamics simulations combined with the stochastic reaction model. The coupling effect of the initiator density and the grafting surface curvature is mainly investigated. The confinement degree greatly changes with the grafting surface curvature, thus the initiation efficiency, the grafted chain polydispersity, as well as the chain mass distribution show great dependence on the surface curvature. The results reveal that preparing the nanoparticle with desired size (i.e., grafting surface curvature) is crucial for control of the grafted chain polydispersity and even its dispersion in the polymer matrix. Our studies shed light on better design of grafted nanoparticles and better control of dispersion in polymer matrices for improving the performance of polymer nanocomposite materials.