High Performance Capabilities Research Articles

GACDNet:Mapping winter wheat by generative adversarial cross-domain networks with transformer integration for zero-sample extraction

Accurate extraction of winter wheat area is essential for wheat yield estimation. Remotely sensed images have limited coverage, are taken at different times, from different angles and in different geographical areas, and the spectral information of the same feature varies from image to image. Machine learning methods are very popular in extracting winter wheat planting area, there are differences between different images, namely the distribution differences between source domain data and target domain data, the results obtained by these methods directly applied to other areas are not satisfactory. To achieve cross regional extraction of winter wheat area, we propose generative adversarial cross-domain networks for image classification under the same image type. The cross-domain network proposed comprises a generative network and a feature extractor. The generative network creates diverse and acceptable samples from the initial input data, constrained by the contrast loss. The feature extractor converts the initial input data and the generated data into a high-level representation. Two time series datasets were constructed and a series of experiments were conducted on these datasets, obtaining high-precision and optimal classification results. Sample training was performed using 70%, 50%, 30% and 10% of the data on both datasets. The study results demonstrated the high performance and cross-domain capability of our network. Furthermore, the winter wheat cultivation area of the entire Zhoukou city was extracted, yielding results with an accuracy of 94.23%. Extensive experiments and applications on the dataset demonstrate that our method is feasible and reliable.

Semi-supervised multi-modal medical image segmentation with unified translation

The two major challenges to deep-learning-based medical image segmentation are multi-modality and a lack of expert annotations. Existing semi-supervised segmentation models can mitigate the problem of insufficient annotations by utilizing a small amount of labeled data. However, most of these models are limited to single-modal data and cannot exploit the complementary information from multi-modal medical images. A few semi-supervised multi-modal models have been proposed recently, but they have rigid structures and require additional training steps for each modality. In this work, we propose a novel flexible method, semi-supervised multi-modal medical image segmentation with unified translation (SMSUT), and a unique semi-supervised procedure that can leverage multi-modal information to improve the semi-supervised segmentation performance. Our architecture capitalizes on unified translation to extract complementary information from multi-modal data which compels the network to focus on the disparities and salient features among each modality. Furthermore, we impose constraints on the model at both pixel and feature levels, to cope with the lack of annotation information and the diverse representations within semi-supervised multi-modal data. We introduce a novel training procedure tailored for semi-supervised multi-modal medical image analysis, by integrating the concept of conditional translation. Our method has a remarkable ability for seamless adaptation to varying numbers of distinct modalities in the training data. Experiments show that our model exceeds the semi-supervised segmentation counterparts in the public datasets which proves our network’s high-performance capabilities and the transferability of our proposed method. The code of our method will be openly available at https://github.com/Sue1347/SMSUT-MedicalImgSegmentation.

Investigating the potential of Fourier transform mid-infrared spectroscopy combined with chemometrics for detecting camel’s milk adulteration

This study explores the effectiveness of mid-infrared (MIR) spectroscopy in combination with chemometrics as an alternative to sensory analysis for detecting camel’s milk adulteration with cow’s milk. A paired comparison test involving various concentrations of adulterants was initially conducted to assess consumers’ ability to detect such adulteration. The analysis successfully classified samples into adulterated and authentic camel’s milk using principal component (PC) analysis and hierarchical cluster analysis. Moreover, the application of partial least squares regression and PC regression calibration models demonstrated high-performance capabilities in revealing the level of adulteration. These findings highlight the potential of MIR spectroscopy combined with chemometrics for the authentication of camel’s milk.

Advancing defense capabilities through integration of electro-optical systems and computer vision technologies

This paper presents a comprehensive study on the integration of cutting-edge technologies to enhance defense capabilities, focusing on object detection, tracking, and distance measurement tasks. Also, the integration of IMX219-77 cameras and Nvidia Jetson Nano for object detection, tracking, and distance measurement in electro-optical systems is proposed. The decision to combine these components was based on their individual strengths: the high resolution and wide field of view of the IMX219-77 cameras, and the low-power, high-performance capabilities of the Jetson Nano. Software-wise, OpenCV and GStreamer were chosen for their widespread use and capabilities in computer vision and multimedia processing tasks, respectively. The MOSSE algorithm was selected for object tracking due to its speed, efficiency, and resilience to lighting changes. Additionally, Depth Estimation (Stereo Vision) was employed for distance measurement by analyzing images captured from a pair of cameras. The proposed algorithm involves transmitting images from the cameras to the Jetson Nano, where OpenCV and GStreamer are utilized to process the data. The MOSSE algorithm is employed for object detection and tracking, while Depth Estimation generates depth maps to measure distances between objects and the camera. The study's results demonstrate accurate distance measurements with a margin of error ranging from 0 to 2 mm, falling below the acceptable threshold mentioned in academic literature. This aligns with similar studies in the field, showcasing the efficacy of the proposed integration for object detection, tracking, and distance measurement in electro-optical systems.

Ultra low RIN, low threshold AlGaInAs/InP BH-DFB laser

This study presents a comparative analysis of DFB lasers based on AlGaInAs/InP heterostructures, exploring the impact of different groove shapes and cavity lengths on laser output characteristics. Leveraging the vertical groove structure, we achieved a remarkable laser output of 90mW at 25°C. Compared to lasers without grooves, this represents a 3.6-fold increase in output power, coupled with an impressively low threshold current of only 4mA. In addition, we conducted a study on the temperature stability of the laser, revealing that even at elevated temperatures of 85°C, the laser maintains an edge mode suppression ratio of over 45dB. This demonstrates the robustness and reliability of the laser over a wide range of operating conditions, further enhancing its suitability for practical applications in diverse environments. Upon packaging the chip into a butterfly shape, we observed a significant reduction in relative intensity noise, with values below -162.8 dB/Hz across the frequency range of 0-40 GHz, while operating at an output power of 3mW. These findings underscore the robustness, reliability, and high-performance capabilities of the developed DFB laser, highlighting its potential for various practical applications.

Shaoxia: a web-based interactive analysis platform for single cell RNA sequencing data

BackgroundIn recent years, Single-cell RNA sequencing (scRNA-seq) is increasingly accessible to researchers of many fields. However, interpreting its data demands proficiency in multiple programming languages and bioinformatic skills, which limited researchers, without such expertise, exploring information from scRNA-seq data. Therefore, there is a tremendous need to develop easy-to-use software, covering all the aspects of scRNA-seq data analysis.ResultsWe proposed a clear analysis framework for scRNA-seq data, which emphasized the fundamental and crucial roles of cell identity annotation, abstracting the analysis process into three stages: upstream analysis, cell annotation and downstream analysis. The framework can equip researchers with a comprehensive understanding of the analysis procedure and facilitate effective data interpretation. Leveraging the developed framework, we engineered Shaoxia, an analysis platform designed to democratize scRNA-seq analysis by accelerating processing through high-performance computing capabilities and offering a user-friendly interface accessible even to wet-lab researchers without programming expertise.ConclusionShaoxia stands as a powerful and user-friendly open-source software for automated scRNA-seq analysis, offering comprehensive functionality for streamlined functional genomics studies. Shaoxia is freely accessible at http://www.shaoxia.cloud, and its source code is publicly available at https://github.com/WiedenWei/shaoxia.

DEEP LEARNING FOR VISUAL RECOGNITION

The design and development of an advanced object detection system are presented in this work, which was guided by a thorough literature review and feasibility assessment. The literature review emphasises how object detection techniques have evolved, highlighting the shift from conventional techniques to deep learning approaches. Important developments are highlighted, such as feature pyramid networks, anchor-based localization, and region-based and single-stage detectors. Furthermore, offered are insights about assessment metrics, transfer learning, and data augmentation. A feasibility study assesses the suggested systems operational, technological, and financial viability and finds that it is highly feasible in each of these areas. The architecture of the system is modular and scalable, including backend services, data management, an object detection engine, and a user interface among its constituent parts. Specifications for both functional and non-functional needs are provided, which direct the development of the system. The development phases, resource allocation, development process, and quality assurance procedures are all outlined in the implementation plan. Through the integration of deep learning techniques, the suggested system seeks to achieve high-performance object identification capabilities that are appropriate for a variety of applications while being scalable, reliable, and user-friendly.

Large-Scale Fabrication of Customized, Tunable, Ultrathin, and Flexible Metamaterial Absorbers Based on Laser-induced Graphene

Developing various 3D graphene microwave absorbers is a hotspot strategy to address the issue of continuous electromagnetic pollution, but remains great challenges in realizing high-performance microwave absorption capability at thin thickness. Herein, a series of flexible metamaterial absorbers (MAs) composed of periodical laser-induced graphene (LIG) surface, dielectric substrate and metallic ground fabric from top to bottom are fabricated. The LIG meta-surface is customized by means of regulating parameters including laser power density (LPD), laser engraving rate (LER) and input patterns to well balance impedance matching and electromagnetic loss. LIG MAs exhibit minimum reflection loss (RL) of −39.7 dB (with effective absorption bandwidth (EAB) of 5.5 GHz) and maximum EBA of 12.1 GHz (with RL of −18.9 dB) at thickness less than 1.1 mm, and the simulated and experimental results are highly consistent. In addition, radar cross section (RCS) simulation of LIG MA with multi-angles endows its potential for prospective applications in in-orbit satellites electromagnetic protection.

Environmental-Friendly Ag2S QD-Multilayer MoSe2 van Der Waals Heterostructure for High-Performance Broadband Photodetection.

Broadband photodetectors have drawn intensive attention owing to their wide application prospects in optical communication, imaging, astronomy, and so on. Two-dimensional transition-metal dichalcogenides (TMDs) are considered as highly potential candidates for photodetection applications, benefiting from their excellent photoelectric properties. However, most of the photodetectors based on TMDs suffer from low performance in the near-infrared (NIR) region due to the weak optical absorption efficiency near their absorption band edge, which severely constrains their usage for broadband optoelectronics. Here, by taking advantage of the high absorption coefficient and environment-friendly property of Ag2S quantum dots (QDs), the hybrid of multilayer MoSe2/Ag2S QDs is demonstrated with a high-performance broadband photodetection capability (532-1270 nm). The favorable energy band alignment of MoSe2/Ag2S QDs facilitates effective separation and collection of photogenerated carriers, and the heterostructure device exhibits significant enhancement of performance compared to the bare MoSe2 device. High responsivity, detectivity, and external quantum efficiency of 25.5 A/W, 1.45 × 1011 Jones, and 1070% are obtained at a low working voltage of 1 V under 980 nm illumination. The responsivity of the device can reach up to 1.2 A/W at 1270 nm wavelength, which is competitive to the commercial NIR photodetectors. Meanwhile, broadband imaging capability is demonstrated. Our work may open up a facile and eco-friendly approach to construct high-performance broadband photodetectors for next-generation compact optoelectronic applications.

Rapid analysis of eucalyptus oil adulteration in Moroccan rosemary essential oil via GC-FID and mid-infrared spectroscopy

Essential Oil (EO) extracted from Rosemary is known for its therapeutic, antifungal, stimulant and antibacterial effects. This study aimed to detect and quantify the adulteration of Rosemary essential oil with different percentages of eucalyptus essential oil, using two analytical techniques: gas chromatography with Flame Ionization Detection (GC-FID) and Fourier Transform Mid-infrared spectroscopy (FT-MIR), combined with chemometric tools such as Principal Component Analysis (PCA), Partial Least Squares regression (PLS-R) and support vector regression (SVR). The use of PCA on the results obtained from GC-FID and FT-MIR indicates the possibility of categorizing the data into two distinct groups: adulterated essential oil and non-adulterated essential oil. However, it is noted that GC-FID can only detect adulteration starting from 40%, while spectroscopy is capable of detecting lower percentages of adulteration. The use of PLS-R and SVR calibration models for adulteration quantification demonstrates high performance capabilities for both techniques (GC-FID and FT-MIR), as indicated by high R2 correlation coefficients indicating good fit, with lower root mean square error (RMSE) values demonstrating predictive accuracy. The results suggest that FT-MIR spectroscopy is preferable to GC-FID for the quantification and discrimination of adulterated essential oils. FT-MIR spectroscopy is considered superior to GC-FID due to its non-destructiveness, speed and lack of sample preparation.

Investigating the Path Dependent Aging Behavior of Nickel Cobalt Aluminum Oxide Cathode Batteries During High C-rate Cycling Conditions

Nickel-rich cathode material batteries such as the nickel cobalt aluminum oxide with graphite anode are common in the automotive sector due to their high energy density and power performance capabilities, however, they are limited in lifetime performance. Accelerated degradation in specific operation conditions, such as high C-rates are known to significantly contribute to this issue. To better understand a chronological dependence of these high-stress conditions occurring during battery lifetime, a path dependent aging study is performed using profiles inspired by high-performance race operation. The focus on C-rates >1C during charging and discharging using real-world inspired profiles- known to induce particle cracking or lithium plating- addresses the need to understand aging behavior in less studied, but highly relevant high-performance automotive context. The aging conditions were tailored to induce specific degradation modes to better identify potential degradation mode coupling. Using differential voltage and capacity analysis techniques, degradation modes were identified which likely contributed to path dependent aging behavior, including loss of lithium inventory and loss of active material in the positive electrode. Two critical paths were identified in this study where later high C-rate conditions in the form of fast-charging at temperatures of 25 °C resulted in accelerated aging and where loss of lithium inventory induced stresses in the positive electrode, accelerating loss of active material.

Detecting possible pairs of materials for composites using a material word co-occurrence network.

Composite materials are popular because of their high performance capabilities, but new material development is time-consuming. To accelerate this process, researchers studying material informatics, an academic discipline combining computational science and material science, have developed less time-consuming approaches for predicting possible material combinations. However, these processes remain problematic because some materials are not suited for them. The limitations of specific candidates for new composites may cause potential new material pairs to be overlooked. To solve this problem, we developed a new method to predict possible composite material pairs by considering more materials than previous techniques. We predicted possible material pairs by conducting link predictions of material word co-occurrence networks while assuming that co-occurring material word pairs in scientific papers on composites were reported as composite materials. As a result, we succeeded in predicting the co-occurrence of material words with high specificity. Nodes tended to link to many other words, generating new links in the created co-occurrence material word network; notably, the number of material words co-occurring with graphene increased rapidly. This phenomenon confirmed that graphene is an attractive composite component. We expect our method to contribute to the accelerated development of new composite materials.

Use of high-performance polymeric materials in customized low-cost robotic grippers for biomechatronic applications: experimental and analytical research

Advancements in materials science and 3D printing technologies have opened up new avenues for developing low-cost robotic grippers with high-performance capabilities, making them suitable for various biomechatronic applications. In this research, it has been explored the utilization of high-performance polymer materials, such as Polyetherketoneketone (PEKK), Polyethylene Terephthalate Glycol (PET-G) and MED 857 (DraftWhite), in the designing and developing of customized robotic grippers. The primary focus of made analyses was oriented on materials characterization, both experimentally and analytically. Computer-Aided Engineering (CAE) methods were employed to simulate bending experiments, allowing for a comprehensive analysis of the mechanical behavior of the selected materials. These simulations were validated through physical bending experiments using samples fabricated via 3D printing technologies, including Fused Filament Fabrication (FFF) for PET-G and PEKK, as well as Jetted Photopolymer (PolyJet) technology employing UV Resin for MED 857. The findings of this research provided advantages of utilizing advanced materials like PEKK in low-cost robotic grippers for biomechatronic applications. The experimental and analytical approaches offer valuable insights into material selection, design optimization, and the development of cost-effective high-performing robotic systems with a wide range of applications in the field of biomechatronics.

Structure-exploiting interior-point solver for high-dimensional entropy-sparsified regression learning

The solution of high-dimensional nonlinear regression problems through standard machine learning approaches often relies on first-order information, due to the numerical and memory challenges arising from the computation of the Hessian matrix and of the higher-order derivatives. While this scenario seems not favorable to second-order methods, here we show that an efficient and modular structure-exploiting interior-point solver can be successfully applied to the recently introduced class of entropy-based methods for regression learning. Specifically, by exploiting the favorable structure of the problem and of the Hessian matrix, we suggest a robust solution strategy based on explicit low-rank updates combined with an iterative Symmetric Quasi-Minimal Residual (SQMR) algorithm to solve the underlying system of linear equations. The results show that the proposed structure-exploiting solver – which relies on the hybrid parallelism and distributed-memory computing paradigm – allows a significant solution time speed-up with respect to a naive solution strategy. Furthermore, through an adequate use of the Message Passing Interface (MPI) and of Open Multi-Processing (OpenMP), the proposed solver enables the solution of large-scale problems on high-performance computing architectures consisting of thousands of compute nodes. The accompanying detailed convergence and performance analyses demonstrate both numerical robustness and high-performance capabilities for increasingly high-dimensional problems.

Prediction of cholesterol content in gallstones by FTIR spectroscopy coupled with chemometric tools

Gallstones are stone-like deposits that most often develop inside the human gallbladder. In this study, the quantification of cholesterol gallstones was monitored using the Fourier transform mid-infrared (FT-MIR) spectroscopy technique and chemometrics tools. To detect the presence of outliers in the spectral data before developing a prediction model, principal component analysis (PCA) was utilized for qualitative analysis. The application of the partial least square (PLS) calibration models for the quantification of cholesterol demonstrated high-performance capabilities, as indicated by high values of correlation coefficients R2 greater than 0.999, and lower values of root mean square error (RMSE) less than 0.493. According to our results, FT-MIR spectroscopy combined with chemometrics approaches can be used successfully as a simple, quick, and non-destructive method for the quantification of cholesterol in biliary lithiasis.

Chlorella vulgaris Harvesting: Chemical Flocculation with Chitosan, Aluminum Sulfate, and Ferric Sulfate

Chlorella vulgaris, a widely cultivated microalgae with diverse commercial applications, faces challenges in economic viability and scalability during the crucial biomass harvesting step. This study explores chemical flocculation followed by sedimentation as a cost-effective solution. Optimization was performed for three flocculants (chitosan, aluminum sulfate, and ferric sulfate), with experiments determining optimal pH and dosage ranges (10–200 mg·L−1). A 24-full factorial design optimized flocculant dosage, settling time, rapid mixing time, and slow mixing time, analyzing their effects on harvesting efficiency through empirical models. The optimal dosage ranges were 50–200 mg·L−1 for aluminum sulfate and 150–200 mg·L−1 for ferric sulfate at pH 9, and 10–50 mg·L−1 for chitosan at pH 5. Empirical models exhibited high fitting performance (R2 > 95%) and predictive capability (predicted R2 > 96%). All flocculants demonstrated high efficiencies (98.4–99.5%), with inorganic types requiring fast and slow mixing phases, while chitosan achieved optimal results without the need for both mixing phases, suggesting potential industrial advantages in time and energy efficiency for microalgae harvesting.

Spatial and temporal variations of surface background ozone in China analyzed with the grid-stretching capability of GEOS-Chem High Performance

Surface background ozone, defined as the ozone in the absence of domestic anthropogenic emissions, is important for developing emission reduction strategies. Here we apply the recently developed GEOS-Chem High Performance (GCHP) global atmospheric chemistry model with ∼0.5° stretched resolution over China to understand the sources of Chinese background ozone (CNB) in the metric of daily maximum 8 h average (MDA8) and to identify the drivers of its interannual variability (IAV) from 2015 to 2019. The GCHP ozone simulations over China are evaluated with an ensemble of surface and aircraft measurements. The five-year national-mean CNB ozone is estimated as 37.9 ppbv, with a spatially west-to-southeast downward gradient (55 to 25 ppbv) and a summer peak (42.5 ppbv). High background levels in western China are due to abundant transport from the free troposphere and adjacent foreign regions, while in eastern China, domestic formation from surface natural precursors is also important. We find greater importance of soil nitric oxides (NOx) than biogenic volatile organic compound emissions to CNB ozone in summer (6.4 vs. 3.9 ppbv), as ozone formation becomes increasingly NOx-sensitive when suppressing anthropogenic emissions. The percentage of daily CNB ozone to total surface ozone generally decreases with increasing daily total ozone, indicating an increased contribution of domestic anthropogenic emissions on polluted days. CNB ozone shows the largest IAV in summer, with standard deviations (seasonal means) of ∼5 ppbv over Qinghai–Tibet Plateau (QTP) and >3.5 ppbv in eastern China. CNB values in QTP are strongly correlated with horizontal circulation anomalies in the middle troposphere, while soil NOx emissions largely drive the IAV in the east. El Nino can inhibit CNB ozone formation in Southeast China by increased precipitation and lower temperature locally in spring, but enhance CNB in Southwest China through increased biomass burning emissions in Southeast Asia.

HW/SW co-design on embedded SoC FPGA for star tracking optimization in space applications

Star trackers are crucial for satellite orientation. Improving their efficiency via reconfigurable COTS HW accommodates NewSpace missions. The current work considers SoC FPGAs to leverage both increased reprogramming and high-performance capabilities. Based on a custom sensor+FPGA system, we develop and optimize the algorithmic chain of star tracking by focusing on the acceleration of the image processing parts. We combine multiple circuit design techniques, such as low-level pipelining, word-length optimization, HW/SW co-processing, and parametric HLS+HDL coding, to fine-tune our implementation on Zynq-7020 FPGA when using real and synthetic input data. Overall, with 4-MPixel images, we achieve more than 24 FPS throughput by accelerating >95% of the computation by 8.9×\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\ imes$$\\end{document}, at system level, while preserving the original SW accuracy and meeting the real-time requirements of the application.

Data-driven coordinated attention deep learning for high-fidelity brain imaging denoising and inpainting.

Deep learning offers promise in enhancing low-quality images by addressing weak fluorescence signals, especially in deep in vivo mouse brain imaging. However, current methods struggle with photon scarcity and noise within in vivo deep mouse brains, and often neglecting tissue preservation. In this study, we propose an innovative in vivo cortical fluorescence image restoration approach, combining signal enhancement, denoising, and inpainting. We curated a deep brain cortical image dataset and developed a novel deep brain coordinate attention restoration network (DeepCAR), integrating coordinate attention with optimized residual networks. Our method swiftly and accurately restores deep cortex images exceeding 800 μm, preserving small-scale tissue structures. It boosts the peak signal-to-noise ratio (PSNR) by 6.94 dB for weak signals and 11.22 dB for large noisy images. Crucially, we validate the effectiveness on external datasets with diverse noise distributions, structural features compared to those in our training data, showcasing real-time high-performance image restoration capabilities.

Nanomaterial-Based Artificial Vision Systems: From Bioinspired Electronic Eyes to In-Sensor Processing Devices.

High-performance robotic vision empowers mobile and humanoid robots to detect and identify their surrounding objects efficiently, which enables them to cooperate with humans and assist human activities. For error-free execution of these robots' tasks, efficient imaging and data processing capabilities are essential, even under diverse and complex environments. However, conventional technologies fall short of meeting the high-standard requirements of robotic vision under such circumstances. Here, we discuss recent progress in artificial vision systems with high-performance imaging and data processing capabilities enabled by distinctive electrical, optical, and mechanical characteristics of nanomaterials surpassing the limitations of traditional silicon technologies. In particular, we focus on nanomaterial-based electronic eyes and in-sensor processing devices inspired by biological eyes and animal visual recognition systems, respectively. We provide perspectives on key nanomaterials, device components, and their functionalities, as well as explain the remaining challenges and future prospects of the artificial vision systems.