Under complex scattering conditions, it is very difficult to capture clear object images hidden behind the media by modelling inverse problem. With regard dynamic media, challenge increases. For solving problem, we propose a new class-specific image reconstruction algorithm. The method based on deep learning classifies blurred according conditions and then recovers media. network used learn mapping relationship between rather than characterizing explicitly or parametrically. 25000 are...

Freeform unobscured multiple-mirror systems have been widely applied in high-precision optical fields due to their high imaging quality and no chromatic aberration central obstruction. However, how design a freeform system with small volume, quality, low manufacturing difficulty is challenging. This study proposes an adaptive generation method for three-mirror volume ease of fabrication based on machine learning, considering the constraints, limitations, efficiency. First, error function...

Imaging under ultra-weak light conditions is affected by Poisson noise heavily. The problem becomes worse if a scattering media present in the optical path. Speckle patterns detected condition carry very little information which makes it difficult to reconstruct image. Off-the-shelf methods are no longer available this condition. In paper, we experimentally demonstrate use of deep learning network images through illumination. weak limitation method analyzed. random detection obtains partial...

Internet of Things (IoT) devices have different operation principles, which weakens the data interoperability. Virtualization is an economic way solving this problem. The data-collected by vendors' sensors-share same computing program encapsulated virtual machine (VM), so that physical-layer difference can be hidden. To eliminate extra cost and long delay transferring VMs to remote cloud, edge device (ED) processes local VMs' requirements in prior. Hence, sustainable strategy for recycling...

The wavemeter is an important instrument for spectrum analysis, widely used in spectral calibration, remote sensing, atomic physics, and high-precision metrology. However, near-infrared (NIR) wavemeters require infrared-sensitive detectors that are expensive less sensitive compared to silicon-based visible light detectors. To circumvent these limitations, we propose NIR speckle based on nonlinear frequency conversion. We combine a scattering medium the deep learning technique invert mapping...

Abstract Exploiting the fundamental features of quantum mechanics, an entanglement-based network offers a promising platform for many dramatic applications such as multi-user cryptography. Nevertheless, implementation large-scale in real-world scenarios remains challenging owing to multiple scattering events complex environment, particularly those frequency-sensitive scatterings that disturb correlation both spatially and temporally. Here, we demonstrate frequency-insensitive spatiotemporal...

Image reconstruction through the opaque medium has great significance in fields of biophotonics, optical imaging, mesoscopic physics, and communications. Previous researches are limited simple linear scattering process. Here, we develop a nonlinear speckle decoder network, which can reconstruct phase information fundamental frequency wave via signal. Further, validate ability our model to recover complex structures by using MNIST CIFAR data sets, respectively. We then show that is able...

Computational imaging enables spatial information retrieval of objects with the use single-pixel detectors. By combining measurements and computational methods, it is possible to reconstruct images in a variety situations that are challenging or impossible traditional multi-pixel cameras. However, these systems typically suffer from significant loss quality due various noises when measurement conditions single-photon detecting, undersampling complicated. Here, we provide an unsupervised deep...

Imaging objects hidden behind an opaque shelter provides a crucial advantage when physically going around the obstacle is impossible or dangerous. Previous methods have demonstrated that possible to reconstruct image of target from view. However, these enable reconstruction by using reflected light wall which may not be feasible in wild. Compared with wall, “plug and play” scattering medium more naturally artificially accessible, such as smog fogs. Here, we introduce scattering-assisted...

Photon-limited imaging has significant application under extreme environments, in which the photon efficiency is an important parameter. In this paper, we investigate of computational ghost with single-photon detection. By exploiting Poisson statistical characteristics counting, relationships between and signal-to-noise ratio retrieved image are theoretically obtained, both closely related to distribution target spatial modulation framework. To verify availability built theoretical model for...

With advancements in freeform surface design and manufacturing, applying surfaces to space cameras is an effective method further enhance imaging quality. However, existing methods for rarely consider quality, system size, manufacturing constraints simultaneously during design. This study proposes adaptive long-focal-length off-axis reflective with lightweight primary/three mirror integration, which balances size ease of while ensuring good First, a configuration the structural parameter...

Typically, the off-axis multi-reflection optical system is assembled by multiple mirrors, which are designed and machined individually due to lack of universal mathematical analytic design technology monolithic manufacturing based on a unified benchmark. It will seriously affect quality overall performance system. In this paper, novel method for systems proposed. An analytical calculation model automatic quantitative under constraints established. The single-point diamond turning adaptive...

Existing design methods for off-axis reflective optical systems have certain limitations, such as complicated processes, lengthy calculations and experience dependence. This study establishes an integrated generation method the initial system of proposes it a solution to aforementioned limitations improvement quality efficiency systems. First, tilt matrix displacement are introduced consider changes in shapes space poses mirrors, simplified mathematical modeling arbitrary pose mirrors is...

We experimentally reconstruct high-quality images through dynamic media using as little ∼0.4 photons per pixel by proposed scalable method, which provides guidance on the real-time photon-limited time-varying scattering imaging applications such in vivo bioimaging.