High-fidelity Images Research Articles

NAPS: Non-adversarial polynomial synthesis

Abstract Generative Adversarial Nets (GANs) are currently the dominant model for high fidelity image synthesis. GANs suffer from two major drawbacks: complicated dynamics and the requirement for an auxiliary network for training (discriminator). However, if we train a decoder-only network we circumvent both drawbacks. To achieve that, the decoder should capture high-order correlations that exist between the variables. We demonstrate this is possible by designing a high-order polynomial generator using tensorial factors. We implement two variants of the model, which we call NAPS. We experiment with both MNIST and CelebA and showcase that our model captures the data distribution and synthesizes new images with significantly less parameters than the corresponding baseline.

Autoferry Gemini: a real-time simulation platform for electromagnetic radiation sensors on autonomous ships

Testing that ships are compliant to specified safety requirements have traditionally relied on real world data, which is not scalable and limited to testable scenarios due to financial and ethical reasons. Low fidelity simulations have been used to counteract some of these problems, which is sufficient for emulating simpler systems such as radar detectors, but not for testing complex systems as found in computer vision. In the automotive industry the use of game engines have shown to be a great testing platform due to their customizability, and combination of real-time physics with computer graphics to create large volumes of high fidelity images. In the work presented here, development of an open-source maritime platform named Autoferry Gemini based on the Unity game engine is used to simulate sensors in real-time. Utilizing simulated optics and general purpose GPU programs, the render pipeline is capable of modeling lidar, radar, visible-light and infrared camera sensors simultaneously. Results from visible-light cameras and lidar have already proven to satisfy other research activities on sensor fusion for autonomous ship technology. Infrared cameras motivates further research in gathering empirical data, while GPU algorithms have made it possible to simulate 3D radar models and multiple lidar types in real-time.

Galaxy and Mass Assembly: A Comparison between Galaxy–Galaxy Lens Searches in KiDS/GAMA

Abstract Strong gravitational lenses are a rare and instructive type of astronomical object. Identification has long relied on serendipity, but different strategies—such as mixed spectroscopy of multiple galaxies along the line of sight, machine-learning algorithms, and citizen science—have been employed to identify these objects as new imaging surveys become available. We report on the comparison between spectroscopic, machine-learning, and citizen-science identification of galaxy–galaxy lens candidates from independently constructed lens catalogs in the common survey area of the equatorial fields of the Galaxy and Mass Assembly survey. In these, we have the opportunity to compare high completeness spectroscopic identifications against high-fidelity imaging from the Kilo Degree Survey used for both machine-learning and citizen-science lens searches. We find that the three methods—spectroscopy, machine learning, and citizen science—identify 47, 47, and 13 candidates, respectively, in the 180 square degrees surveyed. These identifications barely overlap, with only two identified by both citizen science and machine learning. We have traced this discrepancy to inherent differences in the selection functions of each of the three methods, either within their parent samples (i.e., citizen science focuses on low redshift) or inherent to the method (i.e., machine learning is limited by its training sample and prefers well-separated features, while spectroscopy requires sufficient flux from lensed features to lie within the fiber). These differences manifest as separate samples in estimated Einstein radius, lens stellar mass, and lens redshift. The combined sample implies a lens candidate sky density of ∼0.59 deg−2 and can inform the construction of a training set spanning a wider mass–redshift space. A combined approach and refinement of automated searches would result in a more complete sample of galaxy–galaxy lens candidates for future surveys.

A Comprehensive Study of Reversible Data Hiding (RDH) Schemes Based on Pixel Value Ordering (PVO)

Reversible data hiding (RDH) plays a significant role in the field of information security which allows safe transmission of secret data inside a cover media. Additionally, RDH allows extraction of secret data and lossless recovery of the cover media at the decoding end which makes it applicable for the applications like medical imaging and military services. In recent years, variety of RDH techniques such as lossless compression, histogram shifting, difference expansion, prediction error expansion and pixel value ordering (PVO) have been proposed. Among all of them, PVO is one of the most researched RDH techniques as it provides high-fidelity images with decent data hiding capacity. So, to thoroughly understand the evolution of PVO based RDH schemes along with their advantages and disadvantages, this paper presents a comprehensive study of all the existing PVO based RDH schemes with an aim to provide research guidance to the researchers who are interested to work in the field of PVO based reversible data hiding. For this, the working of PVO based schemes is discussed and explained by using suitable examples. Additionally, the challenges being faced by reversible data hiding technique for increasing the data hiding capacity along with image quality, are comprehensively presented. Furthermore, comparative analysis of different PVO-based RDH scheme (after categorizing them based on their working style) using two key parameters—embedding capacity and marked image quality, is presented.

Zwitterionic AIEgens: Rational Molecular Design for NIR‐II Fluorescence Imaging‐Guided Synergistic Phototherapy

AbstractFluorescence imaging in the second near‐infrared region (NIR‐II) can penetrate tissue at centimeter depths and obtain high image fidelity. However, facile synthesis of small‐molecule fluorescent photosensitizers for efficient NIR‐II fluorescence imaging as well as photodynamic and photothermal combinatorial therapies is still a challenging task. Herein, a rational design and facile synthesis protocol are reported for a series of novel NIR‐emissive zwitterionic luminogens with aggregation‐induced emission (AIE) features for cancer phototheranostics. Consistent with the intrinsic features including long emission wavelength, effective reactive oxygen species generation, and excellent photothermal conversion efficiency (35.76%), in vitro and in vivo evaluation show that one of these presented AIE luminogens provides excellent performance in NIR‐II fluorescence imaging‐guided synergistic phototherapy against cancer.

Edge-aware filtering with Siamese neural networks

Recent demands to reduce turnaround times and expedite investment decisions in seismic exploration have invited new ways to process and interpret seismic data. Among these ways is a more integrated collaboration between seismic processors and geologist interpreters aiming to build preliminary geologic models for early business impact. A key aspect has been quick and streamlined delivery of clean high-fidelity 3D seismic images via postmigration filtering capabilities. We present a machine learning-based example of such a capability built on recent advances in deep learning systems. In particular, we leverage the power of Siamese neural networks, a new class of neural networks that is powerful at learning discriminative features. Our novel adaptation, edge-aware filtering, employs a deep Siamese network that ranks similarity between seismic image patches. Once the network is trained, we capitalize on the learned features and self-similarity property of seismic images to achieve within-image stacking power endowed with edge awareness. The method generalizes well to new data sets due to the few-shot learning ability of Siamese networks. Furthermore, the learning-based framework can be extended to a variety of noise types in 3D seismic data. Using a convolutional architecture, we demonstrate on three field data sets that the learned representations lead to superior filtering performance compared to structure-oriented filtering. We examine both filtering quality and ease of application in our analysis. Then, we discuss the potential of edge-aware filtering as a data conditioning tool for rapid structural interpretation.

Reaching thermal noise at ultra-low radio frequencies

Context. Ultra-low frequency observations (< 100 MHz) are particularly challenging because they are usually performed in a low signal-to-noise ratio regime due to the high sky temperature and because of ionospheric disturbances whose effects are inversely proportional to the observing frequency. Nonetheless, these observations are crucial for studying the emission from low-energy populations of cosmic rays. Aims. We aim to obtain the first thermal-noise limited (∼1.5 mJy beam−1) deep continuum radio map using the Low Frequency Array’s Low Band Antenna (LOFAR LBA) system. Our demonstration observation targeted the galaxy cluster RX J0603.3+4214 (known as the Toothbrush cluster). We used the resulting ultra-low frequency (39–78 MHz) image to study cosmic-ray acceleration and evolution in the post shock region considering the presence of a radio halo. Methods. We describe the data reduction we used to calibrate LOFAR LBA observations. The resulting image was combined with observations at higher frequencies (LOFAR 150 MHz and VLA 1500 MHz) to extract spectral information. Results. We obtained the first thermal-noise limited image from an observation carried out with the LOFAR LBA system using all Dutch stations at a central frequency of 58 MHz. With eight hours of data, we reached an rms noise of 1.3 mJy beam−1 at a resolution of 18″ × 11″. Conclusions. The procedure we developed is an important step towards routine high-fidelity imaging with the LOFAR LBA. The analysis of the radio spectra shows that the radio relic extends to distances of 800 kpc downstream from the shock front, larger than what is allowed by electron cooling time. Furthermore, the shock wave started accelerating electrons already at a projected distance of < 300 kpc from the crossing point of the two clusters. These results may be explained by electrons being re-accelerated downstream by background turbulence, possibly combined with projection effects with respect to the radio halo.

Deep Optimal Filter Responses for Multi-Spectral Imaging

Spectral recovery from measured camera signals based on deep learning lead to significant advancements of the potential reconstruction quality. However, most deep learning based approaches only consider RGB cameras and are targeting object classification in particular or remote sensing in general as their final application. Within this work, we analyze the influence of a joint filter optimization and spectral recovery for multi-spectral image acquisition with the underlying goal of capturing high-fidelity color images. An evaluation on the influence of the total camera channel count on the reproduction quality is provided. Finally, a possible normalization of spectral data is discussed.

Two-step system for image receiving in OAM-SK-FSO link.

Turbulence resistance is a significant research area for orbital angular momentum shift keying-based free space optical communication (OAM-SK-FSO). We put forward a two-step combinational system to receive high fidelity image data from the atmospheric turbulence (AT) channels. Firstly, the AT-detector-based multi-CNN (ATDM-CNN) demodulator is proposed which is very different from the traditional single-CNN (S-CNN) demodulator. The AT detector detects the AT strength and then an AT-determined CNN-based demodulator is activated to recognize the incident OAM modes. Sufficient numeral simulations compare the recognition rates of ATDM-CNN and S-CNN. The results indicate a tremendous improvement owing to the ATDM-CNN demodulator. Base on the ATDM-CNN's significant advantage in OAM recognition, the significant optimization of image data quality is possible in the further correction. As an option, the residual information errors are corrected by jointly using the rank-order adaptive median filter (RAMF) and the very-deep super-resolution (VDSR) network with minor information loss in severe ATs. The data increase resulting from RAMF-VDSR is tested. In conclusion, the proposed two-step system can provide a much higher quality of receiving image data in the OAM-SK-FSO link.

Design of Neutron Microscopes Equipped with Wolter Mirror Condenser and Objective Optics for High-Fidelity Imaging and Beam Transport.

We present and compare the designs of three types of neutron microscopes for high-resolution neutron imaging. Like optical microscopes, and unlike standard neutron imaging instruments, these microscopes have both condenser and image-forming objective optics. The optics are glancing-incidence axisymmetric mirrors and therefore suitable for polychromatic neutron beams. The mirrors are designed to provide a magnification of 10 to achieve a spatial resolution of better than 10 m. The resolution of the microscopes is determined by the mirrors rather than by the L/Dratio as in conventional pinhole imaging, leading to possible dramatic improvements in the signal rate. We predict the increase in the signal rate by at least two orders of magnitude for very high-resolution imaging, which is always flux limited. Furthermore, in contrast to pinhole imaging, in the microscope, the samples are placed far from the detector to allow for a bulky sample environment without sacrificing spatial resolution.

Observing protoplanetary discs with the Square Kilometre Array – I. Characterizing pebble substructure caused by forming planets

ABSTRACT High-angular-resolution observations of discs at mm wavelengths (on scales of a few au) are now commonplace, but there is a current lack of a comparable angular resolution for observations at cm wavelengths. This presents a significant barrier to improving our understanding of planet formation, in particular how dust grains grow from mm to cm sizes. In this paper, we examine the ability of the Square Kilometre Array (SKA) to observe dust substructure in a young, planet-forming disc at cm wavelengths. We use dusty hydrodynamics and continuum radiative transfer to predict the distribution and emission of 1-cm dust grains (or pebbles) within the disc, and simulate continuum observations with the current SKA1-MID design baseline at frequencies of 12.5 GHz (Band 5b, ∼2.4 cm) on 5–10 au scales. The SKA will provide high-fidelity observations of the cm dust emission substructure in discs for integration times totalling hundreds of hours. Radial structure can be obtained at a sufficient resolution and S/N from shorter (tens of hours) integration times by azimuthal averaging in the image plane. By modelling the intensity distribution directly in the visibility plane, it is possible to recover a similar level of (axisymmetric) structural detail from observations with integration times one to two orders of magnitude lower than required for high-fidelity imaging. Our results demonstrate that SKA1-MID will provide crucial constraints on the distribution and morphology of the raw material for building planets, the pebbles in protoplanetary discs.

Multiband RadioAstron space VLBI imaging of the jet in quasar S5 0836+710

Context. Detailed studies of relativistic jets in active galactic nuclei (AGN) require high-fidelity imaging at the highest possible resolution. This can be achieved using very long baseline interferometry (VLBI) at radio frequencies, combining worldwide (global) VLBI arrays of radio telescopes with a space-borne antenna on board a satellite. Aims. We present multiwavelength images made of the radio emission in the powerful quasar S5 0836+710, obtained using a global VLBI array and the antenna Spektr-R of the RadioAstron mission of the Russian Space Agency, with the goal of studying the internal structure and physics of the relativistic jet in this object. Methods. The RadioAstron observations at wavelengths of 18 cm, 6 cm, and 1.3 cm are part of the Key Science Program for imaging radio emission in strong AGN. The internal structure of the jet is studied by analyzing transverse intensity profiles and modeling the structural patterns developing in the flow. Results. The RadioAstron images reveal a wealth of structural detail in the jet of S5 0836+710 on angular scales ranging from 0.02 mas to 200 mas. Brightness temperatures in excess of 1013 K are measured in the jet, requiring Doppler factors of ≥100 for reconciling them with the inverse Compton limit. Several oscillatory patterns are identified in the ridge line of the jet and can be explained in terms of the Kelvin–Helmholtz (KH) instability. The oscillatory patterns are interpreted as the surface and body wavelengths of the helical mode of the KH instability. The interpretation provides estimates of the jet Mach number and of the ratio of the jet to the ambient density, which are found to be Mj ≈ 12 and η ≈ 0.33. The ratio of the jet to the ambient density should be conservatively considered an upper limit because its estimate relies on approximations.

Single-pixel compressive imaging based on random DoG filtering

As its name implies, compressive sensing aims to bring compression during sampling. However, the deployment of this technique depends on recovering a high fidelity image through a low number of measurements with a simple hardware and fast software. To this end, we introduce an encoding scheme that by filtering the scene acquires information about the image structure. To prepare a set of proposed encoding patterns, at the first step, a filter bank containing a number of Difference of Gaussian (DoG) kernels with different scales is prepared. Then, by randomly selecting the filters from the bank and under-sampling the scene with them at random points, each encoding pattern is constructed. The idea is inspired by the Human Visual System (HVS) that uses a set of size-tuned DoG kernels at each point in the field-of-view. These encoding patterns, which make a set of linearly independent vectors, form the rows of a structured measurement matrix. This matrix allows making relatively well-conditioned dictionaries by different sparsifying bases. The effectiveness of this method is confirmed by simulations and analyses.

Intelligent Nanoprobe: Acid-Responsive Drug Release and In Situ Evaluation of Its Own Therapeutic Effect.

The design of an intelligent nanoprobe capable of intracellular controlled release of apoptosis inducers and subsequent high-fidelity imaging of the drug-induced apoptosis is highly desirable for precise cancer treatment. Herein, we report an intelligent nanoprobe that combined therapeutic and imaging functions in one agent. Briefly, a gold nanoparticle is designed to be conjugated with acid-responsive DNA duplexes (Dox intercalates in this region) and caspase-3-specific cleavable peptides (labeled with fluorophore). We demonstrated that the nanoprobe could efficiently deliver an anticancer drug (Dox) into cancer cells and achieve acid-responsive drug release. Furthermore, the apoptotic process was in situ-monitored by detection of fluorescence through the cleavage of the peptide linker by caspase-3, which is one of the executioner caspases involved in apoptosis. This newly developed nanoprobe could serve as a theranostic agent for targeted responsive chemotherapy and also provide feedback apoptosis imaging of the self-therapeutic effect.

Advanced Fluorescence Imaging Technology in the Near-Infrared-II Window for Biomedical Applications.

Fluorescence imaging has become a fundamental tool for biomedical applications; nevertheless, its intravital imaging capacity in the conventional wavelength range (400-950 nm) has been restricted by its extremely limited tissue penetration. To tackle this challenge, a novel imaging approach using the fluorescence in the second near-infrared window (NIR-II, 1000-1700 nm) has been developed in the past decade to achieve deep penetration and high-fidelity imaging, and thus significant biomedical applications have begun to emerge. In this Perspective, we first examine recent discoveries and challenges in the development of novel NIR-II fluorophores and compatible imaging apparatuses. Subsequently, the recent advances in bioimaging, biosensing, and therapy using such a cutting-edge imaging technique are highlighted. Finally, based on the achievement in the representative studies, we elucidate the main concerns regarding this imaging technique and give some advice and prospects for the development of NIR-II imaging for future biomedical applications.

Coded coherent diffraction imaging with reduced binary modulations and low-dynamic-range detection.

Conventional coherent diffraction imaging (CDI) suffers from inherent phase retrieval ambiguity due to limited intensity-only measurements. Coded illumination with multiple modulations has been introduced to tackle the underdetermination challenge, which however slows down imaging speed. In addition, the required high-dynamic-range acquisition at the Fourier plane is also time consuming. To increase imaging speed, we report an accelerated coded CDI method in this Letter. It requires only three binarized intensity patterns to illuminate the full field, which can be implemented at ∼22kHz using a digital micromirror device. Each diffraction pattern at the Fourier plane is acquired in a single shot without high-dynamic-range synthesis, resulting in three intensity-only images corrupted with underexposed pixels. We develop an adaptive phase retrieval algorithm to adaptively remove the negative influence of underexposure and recover both the object's amplitude and phase. Both simulations and experiments validate that the method enables fast and high-fidelity complex-field imaging.

End-to-end jet classification of quarks and gluons with the CMS Open Data

We describe the construction of novel end-to-end jet image classifiers to discriminate quark- versus gluon-initiated jets using the simulated CMS Open Data. These multi-detector images correspond to true maps of the low-level energy deposits in the detector, giving the classifiers direct access to the maximum recorded event information about the jet, differing fundamentally from conventional jet images constructed from reconstructed particle-level information. Using this approach, we achieve classification performance competitive with current state-of-the-art jet classifiers that are dominated by particle-based algorithms. We find the performance to be driven by the availability of precise spatial information, highlighting the importance of high-fidelity detector images. We then illustrate how end-to-end jet classification techniques can be incorporated into event classification workflows using Quantum Chromodynamics di-quark versus di-gluon events. We conclude with the end-to-end event classification of full detector images, which we find to be robust against the effects of underlying event and pileup outside the jet regions-of-interest.

Techniques for high-fidelity X-ray micro-tomography of additively manufactured metal components

ABSTRACT Additive manufacturing is revolutionising the production of complex-shaped components and finding widespread application in the industries where production volumes are low and part geometries are complex. High-fidelity 3D imaging of additively manufactured metal parts, as the most effective nondestructive evaluation method, is crucial to maintaining the quality of safety-critical components during manufacturing. However, 3D X-ray imaging of large metallic parts is complicated by strong, nonlinear interactions between the sample and the X-ray beam; phenomena such as beam hardening, photon starvation and X-ray scatter are difficult to account for and cause artefacts in the computed 3D tomographic volume. These artefacts can obscure real defects and cause severe inaccuracies in analysis and evaluation. Here we present several techniques developed in the CTLab at the Australian National University that enhance the micro-tomography instruments’ ability to image large, complex metallic components.

Simulating Tissues with 3D-Printed and Castable Materials.

Manufacturing technologies continue to be developed and utilized in medical prototyping, simulations, and imaging phantom production. For radiologic image-guided simulation and instruction, models should ideally have similar imaging characteristics and physical properties to the tissues they replicate. Due to the proliferation of different printing technologies and materials, there is a diverse and broad range of approaches and materials to consider before embarking on a project. Although many printed materials' biomechanical parameters have been reported, no manufacturer includes medical imaging properties that are essential for realistic phantom production. We hypothesize that there are now ample materials available to create high-fidelity imaging anthropomorphic phantoms using 3D printing and casting of common commercially available materials. A material database of radiological, physical, manufacturing, and economic properties for 29 castable and 68 printable materials was generated from samples fabricated by the authors or obtained from the manufacturer and scanned with CT at multiple tube voltages. This is the largest study assessing multiple different parameters associated with 3D printing to date. These data are being made freely available on GitHub, thus affording medical simulation experts access to a database of relevant imaging characteristics of common printable and castable materials. Full data available at: https://github.com/nmcross/Material-Imaging-Characteristics .

Single-pixel sensing with optimal binarized modulation

Conventional high-level sensing techniques require high-fidelity images as input to extract target features. The images are produced by either complex imaging hardware or high-complexity reconstruction algorithms. In this Letter, we propose single-pixel sensing (SPS) that performs high-level sensing directly from a small amount of coupled single-pixel measurements, without the conventional image acquisition and reconstruction process. The technique consists of three steps, including binarized light modulation at ∼22.7kHz refresh rate, single-pixel coupled detection with a wide working spectrum and high signal-to-noise ratio, and end-to-end deep-learning-based decoding that reduces both hardware and software complexity. Also, the binarized modulation patterns are optimized with the decoding network by a two-step training strategy, leading to the least required measurements and optimal sensing accuracy. The effectiveness of SPS is experimentally demonstrated on the classification task of the handwritten MNIST dataset, and 96% classification accuracy at ∼1kHz is achieved. The reported SPS technique is a novel framework for efficient machine intelligence, with data-reduced acquisition and load-relieved processing.