基于YOLOv8-NFMC的带钢表面缺陷检测算法

针对YOLOv8算法在应用于带钢表面缺陷检测时存在漏检和错检等问题,提出了一种改进YOLOv8算法。针对数据集中的小目标的标签,在原损失CIOU的基础上面加入标准化高斯瓦瑟斯坦距离(normalized Gaussian Wasserstein distance,NWD),提升模型对小目标缺陷的检测能力;采用聚焦调制(focal modulation)替换YOLOv8模型的空间池化金字塔(spatial pyramid pooling-fast,SPPF),在轻量化的同时,提高多尺度特征的表达能力;采用移动翻转瓶颈卷积(mobile inverted bottleneck conv,MBConv)替换C2f中的Conv构建新模块C2f-MB,同时使用C2f-MB替换原有的C2f模块,增强特征表达能力和多尺度特征融合能力;在主干部分加入卷积块注意力机制(convolutional block attention module,CBAM)来抑制背景干扰,能更好捕获全局信息,提升了主干部分的特征提取能力。实验结果表明,改进后的YOLOv8算法在计算量下降的同时,mAP@0.5提高了3%,对漏检和错检等问题有明显改善。

High photoelectric conversion efficiency and fast relaxation time of FA_(0.4)MA_(0.6)PbI_(3) applied in ultrafast modulation of terahertz waves

Active control of terahertz(THz)waves is attracting tremendous attentions in terahertz communications and active photonic devices.Perovskite,due to its excellent photoelectric conversion performance and simple manufacturing process,has emerged as a promising candidate for optoelectronic applications.However,the exploration of perovskites in optically controlled THz modulators is still limited.In this work,the photoelectric properties and carrier dynamics of FA_(0.4)MA_(0.6)PbI_(3)perovskite films were investigated by optical pumped terahertz probe(OPTP)system.The ultrafast carrier dynamics reveal that FA_(0.4)MA_(0.6)PbI_(3)thin film exhibits rapid switching and relaxation time within picosecond level,suggesting that FA_(0.4)MA_(0.6)PbI_(3)is an ideal candidate for active THz devices with ultrafast response.Furthermore,as a proof of concept,a FA_(0.4)MA_(0.6)PbI_(3)-based metadevice with integrating plasma-induced transparency(PIT)effect was fabricated to achieve ultrafast modulation of THz wave.The experimental results demonstrated that the switching time of FA_(0.4)MA_(0.6)PbI_(3)-based THz modulator is near to 3.5 ps,and the threshold of optical pump is as low as 12.7μJ cm^(-2).The simulation results attribute the mechanism of ultrafast THz modulation to photo-induced free carriers in the FA_(0.4)MA_(0.6)PbI_(3)layer,which progressively shorten the capacitive gap of PIT resonator.This study not only illuminates the potential of FA_(0.4)MA_(0.6)PbI_(3)in THz modulation,but also contributes to the field of ultrafast photonic devices.

Design of Protograph LDPC-Coded MIMO-VLC Systems with Generalized Spatial Modulation

This paper investigates the bit-interleaved coded generalized spatial modulation(BICGSM) with iterative decoding(BICGSM-ID) for multiple-input multiple-output(MIMO) visible light communications(VLC). In the BICGSM-ID scheme, the information bits conveyed by the signal-domain(SiD) symbols and the spatial-domain(SpD) light emitting diode(LED)-index patterns are coded by a protograph low-density parity-check(P-LDPC) code. Specifically, we propose a signal-domain symbol expanding and re-allocating(SSER) method for constructing a type of novel generalized spatial modulation(GSM) constellations, referred to as SSERGSM constellations, so as to boost the performance of the BICGSM-ID MIMO-VLC systems.Moreover, by applying a modified PEXIT(MPEXIT) algorithm, we further design a family of rate-compatible P-LDPC codes, referred to as enhanced accumulate-repeat-accumulate(EARA) codes,which possess both excellent decoding thresholds and linear-minimum-distance-growth property. Both analysis and simulation results illustrate that the proposed SSERGSM constellations and P-LDPC codes can remarkably improve the convergence and decoding performance of MIMO-VLC systems. Therefore, the proposed P-LDPC-coded SSERGSM-mapped BICGSMID configuration is envisioned as a promising transmission solution to satisfy the high-throughput requirement of MIMO-VLC applications.

Bile acids,gut microbiota,and therapeutic insights in hepatocellular carcinoma

Hepatocellular carcinoma(HCC)is a prevalent and aggressive liver malignancy.The interplay between bile acids(BAs)and the gut microbiota has emerged as a critical factor in HCC development and progression.Under normal conditions,BA metabolism is tightly regulated through a bidirectional interplay between gut microorganisms and BAs.The gut microbiota plays a critical role in BA metabolism,and BAs are endogenous signaling molecules that help maintain liver and intestinal homeostasis.Of note,dysbiotic changes in the gut microbiota during pathogenesis and cancer development can disrupt BA homeostasis,thereby leading to liver inflammation and fibrosis,and ultimately contributing to HCC development.Therefore,understanding the intricate interplay between BAs and the gut microbiota is crucial for elucidating the mechanisms underlying hepatocarcinogenesis.In this review,we comprehensively explore the roles and functions of BA metabolism,with a focus on the interactions between BAs and gut microorganisms in HCC.Additionally,therapeutic strategies targeting BA metabolism and the gut microbiota are discussed,including the use of BA agonists/antagonists,probiotic/prebiotic and dietary interventions,fecal microbiota transplantation,and engineered bacteria.In summary,understanding the complex BA-microbiota crosstalk can provide valuable insights into HCC development and facilitate the development of innovative therapeutic approaches for liver malignancy.

A Novel Multiple DBC-staked units Package to Parallel More Chips for SiC Power Module

Silicon carbide(SiC) power modules play an essential role in the electric vehicle drive system. To improve their performance, reduce their size, and increase production efficiency, this paper proposes a multiple staked direct bonded copper(DBC) unit based power module packaging method to parallel more chips. This method utilizes mutual inductance cancellation effect to reduce parasitic inductance. Because the conduction area in the new package is doubled, the overall area of power module can be reduced. Entire power module is divided into smaller units to enhance manufacture yield, and improve design freedom. This paper provides a detailed design, analysis and fabrication procedure for the proposed package structure. Additionally, this paper offers several feasible solutions for the connection between power terminals and DBC untis. With the structure, 18dies were paralleled for each phase-leg in a econodual size power module. Both simulation and double pulse test results demonstrate that, compared to conventional layouts, the proposed package method has 74.8% smaller parasitic inductance and 34.9% lower footprint.

From VIB‑to VB‑Group Transition Metal Disulfides:Structure Engineering Modulation for Superior Electromagnetic Wave Absorption

The laminated transition metal disulfides(TMDs),which are well known as typical two-dimensional(2D)semiconductive materials,possess a unique layered structure,leading to their wide-spread applications in various fields,such as catalysis,energy storage,sensing,etc.In recent years,a lot of research work on TMDs based functional materials in the fields of electromagnetic wave absorption(EMA)has been carried out.Therefore,it is of great significance to elaborate the influence of TMDs on EMA in time to speed up the application.In this review,recent advances in the development of electromagnetic wave(EMW)absorbers based on TMDs,ranging from the VIB group to the VB group are summarized.Their compositions,microstructures,electronic properties,and synthesis methods are presented in detail.Particularly,the modulation of structure engineering from the aspects of heterostructures,defects,morphologies and phases are systematically summarized,focusing on optimizing impedance matching and increasing dielectric and magnetic losses in the EMA materials with tunable EMW absorption performance.Milestones as well as the challenges are also identified to guide the design of new TMDs based dielectric EMA materials with high performance.

Automatic modulation recognition of radio fuzes using a DR2D-based adaptive denoising method and textural feature extraction

The identification of intercepted radio fuze modulation types is a prerequisite for decision-making in interference systems.However,the electromagnetic environment of modern battlefields is complex,and the signal-to-noise ratio(SNR)of such environments is usually low,which makes it difficult to implement accurate recognition of radio fuzes.To solve the above problem,a radio fuze automatic modulation recognition(AMR)method for low-SNR environments is proposed.First,an adaptive denoising algorithm based on data rearrangement and the two-dimensional(2D)fast Fourier transform(FFT)(DR2D)is used to reduce the noise of the intercepted radio fuze intermediate frequency(IF)signal.Then,the textural features of the denoised IF signal rearranged data matrix are extracted from the statistical indicator vectors of gray-level cooccurrence matrices(GLCMs),and support vector machines(SVMs)are used for classification.The DR2D-based adaptive denoising algorithm achieves an average correlation coefficient of more than 0.76 for ten fuze types under SNRs of-10 d B and above,which is higher than that of other typical algorithms.The trained SVM classification model achieves an average recognition accuracy of more than 96%on seven modulation types and recognition accuracies of more than 94%on each modulation type under SNRs of-12 d B and above,which represents a good AMR performance of radio fuzes under low SNRs.

Modulated-ISRJ rejection using online dictionary learning for synthetic aperture radar imagery

In electromagnetic countermeasures circumstances,synthetic aperture radar(SAR)imagery usually suffers from severe quality degradation from modulated interrupt sampling repeater jamming(MISRJ),which usually owes considerable coherence with the SAR transmission waveform together with periodical modulation patterns.This paper develops an MISRJ suppression algorithm for SAR imagery with online dictionary learning.In the algorithm,the jamming modulation temporal properties are exploited with extracting and sorting MISRJ slices using fast-time autocorrelation.Online dictionary learning is followed to separate real signals from jamming slices.Under the learned representation,time-varying MISRJs are suppressed effectively.Both simulated and real-measured SAR data are also used to confirm advantages in suppressing time-varying MISRJs over traditional methods.

Approach to loss of response to advanced therapies in inflammatory bowel disease

BACKGROUND Remarkable progress over the last decade has equipped clinicians with many options in the treatment of inflammatory bowel disease.Clinicians now have the unique opportunity to provide individualized treatment that can achieve and sustain remission in many patients.However,issues of primary non-response(PNR)and secondary loss of response(SLOR)to non-tumour necrosis factor inhibitor(TNFi)therapies remains a common problem.Specific issues include the choice of optimization of therapy,identifying when dose optimization will recapture response,establishing optimal dose for escalation and when to switch therapy.AIM To explores the issues of PNR and SLOR to non-TNFi therapies.METHODS This review explores the current evidence and literature to elucidate management options in cases of PNR/SLOR.It will also explore potential predictors for response following SLOR/PNR to therapies including the role of therapeutic drug monitoring(TDM).RESULTS In the setting of PNR and loss of response to alpha-beta7-integrin inhibitors and interleukin(IL)-12 and IL-23 inhibitors dose optimization is a reasonable option to capture response.For Janus kinase inhibitors dose optimization can be utilized to recapture response with loss of response.CONCLUSION The role of TDM in the setting of advanced non-TNFi therapies to identify patients who require dose optimization and as a predictor for clinical remission is not yet established and this remains an area that should be addressed in the future.

基于SGP4模型的空间站轨道精度分析

针对在地面站天线对空间站观测任务中,通常基于卫星工具包(satellite tool kit,STK)软件规划出天线对空间站的方位角和俯仰角,实现天线对空间站的自动跟踪.为了保证天线跟踪的准确性和可靠性,需要定期计算出准确的空间站轨道和天线的方位俯仰角,并更新规划任务.因此科学分析与评估空间站两行轨道根数(two line elements,TLE)长期预报精度,对地面站实现空间站的精准跟踪具有重要意义.本文以中国空间站(China Space Station,CSS)梦天实验舱为例,基于TLE数据,利用STK软件提供的简化常规摄动规模型(simplified general perturbations4,SGP4)模型计算空间站轨道以及空间站相对于西安地面站的方位角和俯仰角,并分析不同策略下的精度效果.试验结果表明:在第二天更新空间站的TLE,可以获得较好的轨道结果,从而更好地保障天线的跟踪精度.

改进YOLOv5s的人员跌倒检测算法

针对跌倒检测任务中复杂信息干扰和数据集缺乏导致模型精度不高的问题,设计一种高精度跌倒检测算法,降低模型参数的同时保持各种场景下的鲁棒性。该算法基于YOLOv5s改进,在骨干网络中使用Ghost module和解耦全连接注意力,以较低计算开销提升模型在光线变化、遮挡等干扰环境下的性能。在颈部层使用自适应感受野和空间通道混合注意力,提升神经元对不同尺度特征的适应性,应对人体形变、视角变化等干扰。引入EIoU损失函数,加速收敛提升训练精度。在公开数据集Le2i Fall Detection Dataset和UR Fall Detection上,精确率、召回率、mAP0.5和mAP(0.5:0.95)相比YOLOv5s分别提高4.0%,4.2%,2.9%和4.3%,参数量降低38.6%。该算法在多种应用场景下都保持较高检测精度,参数量较低,满足实际应用场景部署要求。

Identification of banana leaf disease based on KVA and GR-ARNet

Banana is a significant crop,and three banana leaf diseases,including Sigatoka,Cordana and Pestalotiopsis,have the potential to have a serious impact on banana production.Existing studies are insufficient to provide a reliable method for accurately identifying banana leaf diseases.Therefore,this paper proposes a novel method to identify banana leaf diseases.First,a new algorithm called K-scale VisuShrink algorithm(KVA)is proposed to denoise banana leaf images.The proposed algorithm introduces a new decomposition scale K based on the semi-soft and middle course thresholds,the ideal threshold solution is obtained and substituted with the newly established threshold function to obtain a less noisy banana leaf image.Then,this paper proposes a novel network for image identification called Ghost ResNeSt-Attention RReLU-Swish Net(GR-ARNet)based on Resnet50.In this,the Ghost Module is implemented to improve the network's effectiveness in extracting deep feature information on banana leaf diseases and the identification speed;the ResNeSt Module adjusts the weight of each channel,increasing the ability of banana disease feature extraction and effectively reducing the error rate of similar disease identification;the model's computational speed is increased using the hybrid activation function of RReLU and Swish.Our model achieves an average accuracy of 96.98%and a precision of 89.31%applied to 13,021 images,demonstrating that the proposed method can effectively identify banana leaf diseases.

面向石墨电极标签识别的轻量化网络LGSNet

轻量化网络已成为面向工业场景部署的关键技术。为进一步提升Ghost module的特征提取能力并减少参数量,提出了一种改进的S-Ghost瓶颈模块(Small Ghost Bottleneck)。此瓶颈模块采用1×1卷积通道与Ghost module并联的结构,缩减Ghost module的通道数以压缩参数量,并用与之并联的1×1卷积进行通道扩充;在模块输出端引入通道混洗(Channel Shuffle)操作以保证通道间信息交互。实验结果表明,利用该瓶颈结构设计的图像分类网络LGSNet (Light Ghost Networks,LGSNet),其计算量和参数量显著降低,同时网络精度与推理速度未受影响,甚至在一些测试中取得最优,此网络设计满足工业需求。这为面向工业场景的轻量化网络架构设计提供了新的解决方案和思路。

Influence of substrate effect on near-field radiative modulator based on biaxial hyperbolic materials

Relative rotation between the emitter and receiver could effectively modulate the near-field radiative heat transfer(NFRHT)in anisotropic media.Due to the strong in-plane anisotropy,natural hyperbolic materials can be used to construct near-field radiative modulators with excellent modulation effects.However,in practical applications,natural hyperbolic materials need to be deposited on the substrate,and the influence of substrate on modulation effect has not been studied yet.In this work,we investigate the influence of substrate effect on near-field radiative modulator based onα-MoO_(3).The results show that compared to the situation without a substrate,the presence of both lossless and lossy substrate will reduce the modulation contrast(MC)for different film thicknesses.When the real or imaginary component of the substrate permittivity increases,the mismatch of hyperbolic phonon polaritons(HPPs)weakens,resulting in a reduction in MC.By reducing the real and imaginary components of substrate permittivity,the MC can be significantly improved,reaching 4.64 forε_(s)=3 at t=10 nm.This work indicates that choosing a substrate with a smaller permittivity helps to achieve a better modulation effect,and provides guidance for the application of natural hyperbolic materials in the near-field radiative modulator.

An enhanced method for predicting and analysing forest fires using an attention-based CNN model

Prediction,prevention,and control of forest fires are crucial on at all scales.Developing effective fire detection systems can aid in their control.This study proposes a novel CNN(convolutional neural network)using an attention blocks module which combines an attention module with numerous input layers to enhance the performance of neural networks.The suggested model focuses on predicting the damage affected/burned areas due to possible wildfires and evaluating the multilateral interactions between the pertinent factors.The results show the impacts of CNN using attention blocks for feature extraction and to better understand how ecosystems are affected by meteorological factors.For selected meteorological data,RMSE 12.08 and MAE 7.45 values provide higher predictive power for selecting relevant and necessary features to provide optimal performance with less operational and computational costs.These findings show that the suggested strategy is reliable and effective for planning and managing fire-prone regions as well as for predicting forest fire damage.

Physical exercise and traumatic brain injury: is it question of time?

Is it better to be safe than sorry?This Hamletic dilemma has always stimulated medical-scientific debates in numerous fields of biomedicine.And among these,the preventive-therapeutic approach to the treatment of brain trauma is one of the most striking examples.Traumatic brain injury(TBI)is a leading cause of brain damage among young and elderly populations with a very high hospitalization and death rate.TBI is characterized by two pathologically distinct but strictly consequential phases:a first characterized by an immediate and highly variable mechanical dysfunction of the brain tissue,which involves widespread cell death and tissue degeneration,followed by a second phase which can last from days to even years depending on the severity of the TBI and the patient’s pre-existing health status.Secondary processes,including inflammatory phenomena,oxidative stress associated with metabolic,vascular,and neuro-modulatory deficits,are very often responsible for neuro-motor and psychological deficits leading to long-term disabilities(Kaur and Sharma,2018).

A review on the cooling of energy conversion and storage systems using thermoelectric modules

Exploitation of sustainable energy sources requires the use of unique conversion and storage systems,such as solar panels,batteries,fuel cells,and electronic equipment.Thermal load management of these energy conversion and storage systems is one of their challenges and concerns.In this article,the thermal management of these systems using thermoelectric modules is reviewed.The results show that by choosing the right option to remove heat from the hot side of the thermoelectric modules,it will be a suitable local cooling,and the thermoelectric modules increase the power and lifespan of the system by reducing the spot temperature.Thermoelectric modules were effective in reducing panel temperature.They increase the time to reach a temperature above 50℃ in batteries by 3 to 4 times.Also,in their integration with fuel cells,they increase the power density of the fuel cell.

First demonstration of a self-aligned p-channel GaN back gate injection transistor

In this study,we present the development of self-aligned p-channel Ga N back gate injection transistors(SA-BGITs)that exhibit a high ON-state current.This achievement is primarily attributed to the conductivity modulation effect of the 2-D electron gas(2DEG,the back gate)beneath the 2-D hole gas(2DHG)channel.SA-BGITs with a gate length of 1μm have achieved an impressive peak drain current(I_(D,MAX))of 9.9 m A/mm.The fabricated SA-BGITs also possess a threshold voltage of 0.15 V,an exceptionally minimal threshold hysteresis of 0.2 V,a high switching ratio of 10~7,and a reduced ON-resistance(RON)of 548Ω·mm.Additionally,the SA-BGITs exhibit a steep sub-threshold swing(SS)of 173 mV/dec,further highlighting their suitability for integration into Ga N logic circuits.

Attention Mechanism-Based Method for Intrusion Target Recognition in Railway

The detection of foreign object intrusion is crucial for ensuring the safety of railway operations.To address challenges such as low efficiency,suboptimal detection accuracy,and slow detection speed inherent in conventional comprehensive video monitoring systems for railways,a railway foreign object intrusion recognition and detection system is conceived and implemented using edge computing and deep learning technologies.In a bid to raise detection accuracy,the convolutional block attention module(CBAM),including spatial and channel attention modules,is seamlessly integrated into the YOLOv5 model,giving rise to the CBAM-YOLOv5 model.Furthermore,the distance intersection-over-union_non-maximum suppression(DIo U_NMS)algorithm is employed in lieu of the weighted nonmaximum suppression algorithm,resulting in improved detection performance for intrusive targets.To accelerate detection speed,the model undergoes pruning based on the batch normalization(BN)layer,and Tensor RT inference acceleration techniques are employed,culminating in the successful deployment of the algorithm on edge devices.The CBAM-YOLOv5 model exhibits a notable 2.1%enhancement in detection accuracy when evaluated on a selfconstructed railway dataset,achieving 95.0%for mean average precision(m AP).Furthermore,the inference speed on edge devices attains a commendable 15 frame/s.

Two-Layer Attention Feature Pyramid Network for Small Object Detection

Effective small object detection is crucial in various applications including urban intelligent transportation and pedestrian detection.However,small objects are difficult to detect accurately because they contain less information.Many current methods,particularly those based on Feature Pyramid Network(FPN),address this challenge by leveraging multi-scale feature fusion.However,existing FPN-based methods often suffer from inadequate feature fusion due to varying resolutions across different layers,leading to suboptimal small object detection.To address this problem,we propose the Two-layerAttention Feature Pyramid Network(TA-FPN),featuring two key modules:the Two-layer Attention Module(TAM)and the Small Object Detail Enhancement Module(SODEM).TAM uses the attention module to make the network more focused on the semantic information of the object and fuse it to the lower layer,so that each layer contains similar semantic information,to alleviate the problem of small object information being submerged due to semantic gaps between different layers.At the same time,SODEM is introduced to strengthen the local features of the object,suppress background noise,enhance the information details of the small object,and fuse the enhanced features to other feature layers to ensure that each layer is rich in small object information,to improve small object detection accuracy.Our extensive experiments on challenging datasets such as Microsoft Common Objects inContext(MSCOCO)and Pattern Analysis Statistical Modelling and Computational Learning,Visual Object Classes(PASCAL VOC)demonstrate the validity of the proposedmethod.Experimental results show a significant improvement in small object detection accuracy compared to state-of-theart detectors.