Steel surface defects usually exhibit small scales, irregular shapes, blurred boundaries, and cluttered background textures, which makes it difficult to jointly optimize detection accuracy and deployment efficiency. To address these challenges, this paper proposes FRG-DETR, a lightweight improved detector for steel surface defect detection built upon RT-DETR. The proposed method improves the baseline from three aspects: backbone feature extraction, front-end downsampling, and bounding-box regression. First, FDAMBlock replaces the original ResNet-18 backbone to enhance the collaborative modeling of global and local information while reducing network complexity. Second, RLStem is introduced to preserve shallow texture and edge details, thereby improving the detection of small and weak-texture defects. Third, an IGC loss is designed to jointly constrain location, scale, and shape discrepancies between predicted and ground-truth boxes, improving regression precision. On NEU-DET, FRG-DETR achieves 79.7% mAP50, outperforming the RT-DETR baseline by 3.8 points, while reducing the number of parameters to 15.8M and the computational cost to 44.7 GFLOPs. Cross-dataset evaluation on GC10-DET further yields 72.0% mAP50, demonstrating good generalization ability. Qualitative results show that the proposed detector focuses more accurately on defect regions and suppresses irrelevant background responses. Overall, FRG-DETR achieves an effective balance between detection performance and deployment efficiency for steel surface defect detection.

Based on the poor adhesion of water-based UV ink on coated paper, in order to improve the adhesion and curing rate of water-based UV ink on coated paper, this study used polyethylene glycol and dihydroxymethylpropionic acid as raw materials, reacted with isoferone diisocyanate at room temperature to synthesize water-based polyurethane. The waterborne UV acrylic polyurethane ink WPU-g-HEMA was prepared by introducing hydroxyethyl methacrylate (HEMA) into the side chain of waterborne polyurethane. Then glycidyl methacrylate (GMA) reacted with diethylene triamine (DETA) to prepare ink diluents. Secondly, a water-based UV ink with fast drying and curing speed and water-alcohol system was prepared. Water-alcohol soluble WPU-g-HEMA was used as ink prepolymer, different molar ratios of GMA and DETA reaction products were used as ink diluents, and 907 (2,4, 6-triethylbenzoyl triethyltin), ITX (2-isopropyl-9, 10-dioxanthene), BPO (benzoyl peroxide) were used as photoinitiators. The effects of different types and contents of different photoinitiators (907: 0.2 %-3.1 %, ITX: 0.2 %-2.8 %, BPO: 0.2 %-3.0 %), the type and content of waterborne UV acrylic polyurethane ink (30 %-70 %), the number and content of ink diluents with different functional groups (20 %-60 %) on the UV curing rate were studied. The adhesion, contact angle, gloss and wear resistance of ink film on coated paper were tested. The experimental results showed that when the mass fraction of photoinitiator Irgacure907 was 3.1 wt%, the mass fraction of the waterborne UV acrylic polyurethane ink WPU-g-HEMA was 7.5 % (59.7 wt%), and the molar ratio of diluent nDETA:nGMA was 1:5 (the number of functional groups was 5, the mass fraction was 38.9 wt%), waterbased UV polyurethane ink had the best light curing performance. The minimum photocuring time was 0.7 s, the gloss was 50 GU, the minimum mass loss of the ink film was 5 %, the minimum contact angle was 53 degrees, and the adhesion was 98 %. These results showed that the ink had good performance and potential application value.

This paper designs a wearable pulse sensor based on piezoelectric film to address issues such as discomfort, inconvenience, and low accuracy in traditional pulse sensors. The sensor aims to achieve continuous detection of human pulse signals, providing robust support for the prevention and treatment of carwere prepared as the sensor substrate by using the tape-casting method. Conductive electrodes were printed on the film's surface via screen-printing technology, and square- and circular-array sensors were developed by incorporating a mesh shielding layer design for comparative performance evaluation in pulse signal acquisition. Second, to address the low-frequency and weak nature of pulse signals, which are prone to various noise interferences, a precise signal conditioning circuit with amplification and filtering functionalities was designed to acquire high-fidelity and low-noise pulse wave signals. Experimental results demonstrate that the prepared P(VDF-TrFE) film exhibits excellent dielectric, piezoelectric, and ferroelectric properties with a maximum d33 value of -25 pC & centerdot; N-1, enhancing the sensor's ability to rapidly and accurately capture low-frequency pulse signals. The designed flexible pulse sensor conforms well to human skin, meeting wearable and comfort requirements. Among the tested designs, the circular-array sensor detected continuous pulse wave signals with the most physiological characteristic points, exhibiting higher sensitivity and clarity than the square sensor and demonstrating superior detection performance. Additionally, the designed signal conditioning circuit effectively mitigated 50 Hz power frequency and high-frequency noise interferences, successfully amplifying the average peak voltage from 0.069 V to 5.467 V. It displayed a clear and stable pulse waveform while retaining the primary features of the pulse signal, achieving high sensitivity, stability, and accuracy in signal acquisition while suppressing noise. Therefore, the wearable pulse sensor, based on the flexible piezoelectric film, designed in our work effectively detects and captures human pulse wave signals, offering significant potential for applications in medical health monitoring and wearable device research.

JPEG images have become ubiquitous in our daily lives due to their ability to strike a balance between visual quality and file size. Consequently, reversible data hiding (RDH) schemes for JPEG images have emerged. However, existing methods by modifying discrete cosine transform (DCT) coefficients still have room for improvement in terms of distortion reduction and file size preservation. This paper proposes a novel two-dimensional (2D) histogram mapping strategy for RDH in JPEG images. Firstly, the quantized DCT coefficient blocks are sorted based on the count of zero alternating current (AC) coefficients in each block. This enables the estimation of distortion in non-zero AC coefficients at different locations within the block, facilitating the selection of frequency coefficients with minimal distortion for embedding. Additionally, leveraging statistical characteristics, six mapping types with different embedding efficiency are created and the improved 2D histogram mapping method is proposed. Furthermore, The unit distortion-increase ratio (UDIR) is employed as a comprehensive evaluation metric. Extensive experiments are conducted on four representative images and two widely-used image databases to validate the proposed scheme. The results consistently demonstrate superior visual quality, minimal file size increase (FSI), and higher UDIR in comparison to state-of-the-art RDH schemes for JPEG images.

This study presents the design and implementation of a surface plasmon resonance (SPR) sensor in the Kretschmann configuration, employing a gold film deposited on a flexible polydimethylsiloxane (PDMS) substrate as the SPR chip. The refractive-index sensitivity of the SPR sensor was evaluated with sodium chloride solutions of varying concentrations. Optimizing for both sensitivity and detection accuracy, the incident angle was fixed at 13°. The sensor exhibited a sensitivity of 3385.5 nm/RIU. Remarkably, the sensitivity variation was merely 1% after subjecting the sensor chip to 50 bending cycles in both forward and reverse directions. The sensor’s efficacy was further validated through the detection of alcohol content in three different Chinese Baijiu samples, yielding a maximum relative error of 4.04% and a minimum error of 0.17%. Additionally, the sensor was utilized to study the adsorption behavior of glutathione (GSH) on the gold film under varying pH conditions. The findings revealed optimal immediate adsorption at pH = 12, attributed to the complete deprotonation of mercapto groups, facilitating the formation of Au-S bonds with gold atoms. The best film-forming effect was observed at pH = 7, where the interplay of attractive and repulsive forces among different molecular groups led to the gradual extension of the molecular chain, resulting in a thicker molecular film.

In this paper, we propose a spectrally-resolved titanium dioxide waveguide resonant sensor, modified with a perovskite film. The sensor consists of a prism (N-FK51A), gold (Au), hybrid organic–inorganic halide perovskites, and titanium dioxide layers. The thickness of each layer and the incident angle were optimized based on the Fresnel equations. The refractive index sensitivity of the sensor increased by 32.1

Wearable electrochemical sensors, which can track various biomarkers in real-time, is one of the most promising bioanalytical devices in the point of care testing and health management. Sample collection modules, as the important front-end of the sensors, capture and transfer enough sample so as to achieve real-time, precise and accurate detection. Noninvasive and minimally invasive techniques for sample collection are highly desired. However, it is still a key challenge to realize a reliable sample acquisition. Herein, we provide an overview on the progress in the sample collection methods for wearable electrochemical sensors. It covers the conventional and emerging methods/structures including reverse iontophoresis, microneedles, as well as microfluidic chips. The working principle, fabrication materials and manufacturing technologies will be introduced in detail. Finally, the present challenges and the development trend in the future of these approaches are also put forward. image

In the domain of monitoring and conservation of rare and endangered wildlife, traditional object detection methods fail to operate effectively due to overfitting caused by limited image data. To address this issue, this paper proposes a small-sample object detection method that utilizes limited support information to achieve accurate detection of endangered wildlife images. Firstly, we introduce the Feature Pyramid Network (FPN) to adapt to foreground targets at various scales. Secondly, we propose a Coupled Relationship Activation (CRA) module that aggregates features from the support set into a comprehensive feature vector and utilizes it to activate similar regions in the query feature. Furthermore, to balance the differences between support and query features, we incorporate cosine similarity and short-circuit structures before and after the activation process, thereby mitigating the impact of activating dissimilar features erroneously. During the above process, we construct a balanced dataset, CRBAD, which comprises images of both endangered and common animal categories, and employ it to conduct practical testing of the proposed method. Experimental results demonstrate that the proposed model achieves state-of-the-art performance in detecting endangered wildlife images.

The fine-grained image retrieval task is challenging due to small inter-class and large intra-class variations, making it hard for hashing methods to distinguish subtle differences. Traditional algorithms often overlook detail in feature extraction, mapping similar images to different hash codes, while existing methods focus on dense regions and ignore other discriminative areas, lowering accuracy. Additionally, current approaches lack inter-channel correlation modeling, limiting feature aggregation. This paper proposes a hashing method based on salient region localization to capture fine-grained local features and an interactive channel transformation module to model inter-channel relationships. Experiments on four benchmark datasets show significant retrieval accuracy improvements over baselines, enhancing fine-grained image retrieval performance.

Arterial blood pressure (ABP) serves as a pivotal clinical metric in cardiovascular health assessments, with the precise forecasting of continuous blood pressure assuming a critical role in both preventing and treating cardiovascular diseases. This study proposes a novel continuous non-invasive blood pressure prediction model, DSRUnet, based on deep sparse residual U-net combined with improved SE skip connections, which aim to enhance the accuracy of using photoplethysmography (PPG) signals for continuous blood pressure prediction. The model first introduces a sparse residual connection approach for path contraction and expansion, facilitating richer information fusion and feature expansion to better capture subtle variations in the original PPG signals, thereby enhancing the network’s representational capacity and predictive performance and mitigating potential degradation in the network performance. Furthermore, an enhanced SE-GRU module was embedded in the skip connections to model and weight global information using an attention mechanism, capturing the temporal features of the PPG pulse signals through GRU layers to improve the quality of the transferred feature information and reduce redundant feature learning. Finally, a deep supervision mechanism was incorporated into the decoder module to guide the lower-level network to learn effective feature representations, alleviating the problem of gradient vanishing and facilitating effective training of the network. The proposed DSRUnet model was trained and tested on the publicly available UCI-BP dataset, with the average absolute errors for predicting systolic blood pressure (SBP), diastolic blood pressure (DBP), and mean blood pressure (MBP) being 3.36 ± 6.61 mmHg, 2.35 ± 4.54 mmHg, and 2.21 ± 4.36 mmHg, respectively, meeting the standards set by the Association for the Advancement of Medical Instrumentation (AAMI), and achieving Grade A according to the British Hypertension Society (BHS) Standard for SBP and DBP predictions. Through ablation experiments and comparisons with other state-of-the-art methods, the effectiveness of DSRUnet in blood pressure prediction tasks, particularly for SBP, which generally yields poor prediction results, was significantly higher. The experimental results demonstrate that the DSRUnet model can accurately utilize PPG signals for real-time continuous blood pressure prediction and obtain high-quality and high-precision blood pressure prediction waveforms. Due to its non-invasiveness, continuity, and clinical relevance, the model may have significant implications for clinical applications in hospitals and research on wearable devices in daily life.

Molybdenum disulfide (MoS2), as a two-dimensional material, has a potential application prospect in microwave absorption, electromagnetic shielding, energy storage, electrochemistry, and other fields due to its unique structure, excellent electrical conductivity, abundant surface area, and good mechanical strength. In this paper, the microwave absorbing properties of MoS2-based composites are reviewed objectively, and the recent research achievements and progress in dielectric loss type and magnetic loss type composites are summarized. Then, the microwave absorption mechanism of MoS2-based composites is systematically analyzed in terms of dielectric loss and magnetic loss. At the same time, the application of MoS2-based microwave absorbing materials in current social needs is summarized including military radar stealth and civil electronic communication. In addition, the challenges and bottlenecks in the future development of microwave absorber materials are also presented.

A group of (0.65Bi0.5Na0.5–0.35Bi0.2Sr0.7)(Ti1−x,Zrx)O3 (BNBS-(Ti1−x,Zrx)) lead-free energy storage ceramic sheets are prepared by a conventional solid-state sintering method. We find that B-site doping of ZrO2 may minimize the grain size while not change the perovskite structure of BNBS-(Ti1−x,Zrx). As such, BNBS-(Ti1−x,Zrx) possesses the uniform grains and clear grain boundaries, resulting in a high dielectric permittivity (εr) of about 2080 and a low dielectric loss (tanδ) of 0.05 at 100 Hz. Due to the lattice distortion caused by Zr4+ entering the TiO6 octahedral lattice, Tm of BNBS-(Ti1−x,Zrx) decreases with the increase of ZrO2 doping content. Attractively, BNBS-(Ti1−x,Zrx) exhibits relatively slender polarization-electric field (P-E) loops at a high electric field of 100 kV/cm, and BNBS-(Ti0.97,Zr0.03) achieves considerable recycle discharging energy density (Wrec) of 1.47 J/cm3 and high efficiency (η) of 86.94

To realize a piezoelectric sensor array for identifying the location of impact stress, the large-scale (200mm x 200 mm) poly(vinylidene-trifluoride) (P(VDF-TrFE) 75/25mol%) films with thickness range from 20 mu m to 80 mu m are prepared using traditional solution casting method, which exhibits the favorable ferroelectric and electromechanical properties. By using screen printing with silver paste electrode, an orthohexagonal-shaped sensor array containing 19 P(VDF-TrFE) units is obtained. Then the dependence of output voltage of piezoelectric signal on the thickness of film and enforced impulse stress is investigated where an impact force testing machine is employed to simulate the impact of a flying ping pong ball. This piezoelectric sensor exhibits a large signal response of 8.9 V under an impulse stress of 1.5 MPa at an impacting speed of 500 mm/min. After signal acquisition and processing, the output signal wave of main senor and adjacent unit can be clearly distinguished from each other, and this array shows a fast response speed of 2 ms, indicating that this sensor array has a prospect of application in intelligent table tennis racket and other impulse signal monitoring area.

Flexible piezoelectric composite, combining high piezoelectricity of filler and flexibility of polymer, provides a new research idea for developing flexible piezoelectric sensors (FPSs) with high piezoelectric output performance. FPSs based on GR/KNN/P(VDF-TrFE) three-phase composites are fabricated via doping a mass fraction of 15 wt% potassium sodium niobate (KNN) ceramic powder and various contents of graphene (GR) nanosheets into a P(VDF-TrFE) matrix. We find that an appropriate amount of GR is responsible for the enhanced crystallinity and β-phase of P(VDF-TrFE). When the GR content is 0.15 wt%, the three-phase composite film exhibits a dielectric constant (εr) of 20.9 and a quasi-static piezoelectric constant (d33) of -28.4 pC N-1. Under three different test scenarios of a ball drop experiment, a surface of the mouse wheel, and an action of 2.5 MPa external stress, this GR/KNN/P(VDF-TrFE)-based FPS shows high piezoelectric output voltages of 7.4 V, ∼2.0 V, and 15.4 V, respectively. Moreover, the FPS retains its performance even after an extended period of cantilever vibration cycles (2200). Thus, the conductive filler GR is responsible for promoting the energy conversion performance of the piezoelectric polymer, which also provides an application candidate of this GR/KNN/P(VDF-TrFE) film in FPSs.

In the real world, the data distribution often presents a long tail distribution, and the imbalance of data will lead to the model learning bias to the head class. To address the influence of long tail distribution on image classification, this paper proposes a feature channel interactive long tail image classification model based on dual attention. Firstly, the dual attention module is used to capture the autocorrelation and spatial dimension information of the feature map, and the enhanced image is obtained by transformation and class activation map. After that, image preprocessing is performed on the enhanced data set to reduce the over-fitting of the model to the head, and the features that are more conducive to tail classification are obtained through learning. Finally, by interacting with the local channels adjacent to the features, the correlation between the channels is extracted to obtain more robust features. The method achieves good performance on CIFAR10-LT, CIFAR100-LT and ImageNet datasets, which proves the effectiveness of the model.

Image descriptor based on Convolution Neural Network (CNN) has been presented to represent image in recent lots of works, outperforming the traditional feature for image retrieval. In this paper, a new method is proposed to capture real object regions in images for better search. In contrast to the existing method that process image as a whole, our method focuses on searching for approximate object region in an image. The proposed method mainly has the following two aspects: (i) we employ the sliding windows with the multiple scales over the image, to extract corresponding local CNN features. After processing feature, an optimized image representation with several categories of weights has been prepared for image retrieval and re-rank. (ii) On the basis of (i), we construct a new framework called approximate object location to search out the most similar domain, which can be applied in renewed retrieval, in an image for the query. This proposed framework provides a vector consisting of many locals in different scales for an image. According to experiments, our proposed method outperforms most current approaches based on CNN, and excels many previous algorithms based on costly bag-of-words.

目的 基于Transformer架构的网络在图像分类中表现出优异的性能.然而,注意力机制往往只关注图像中的显著性特征,而忽略了其他区域的次级显著信息,基于自注意力机制的Transformer也是如此.为了获取更多的有效信息,从有区别的潜在性特征中学习到更多的可判别特征,提出了一种互补注意多样性特征融合网络(comple-mentary attention diversity feature fusion network,CADF),通过关注次显特征和对通道与空间特征协同编码,以增强特征多样性的注意感知.方法 CADF由潜在性特征模块(potential feature module,PFM)和多样性特征融合模块(diversity feature fusion module,DFFM)组成.PFM模块通过聚合空间与通道中感兴趣区域得到显著性特征,再对特征的显著性进行抑制,以强制网络挖掘潜在性特征,从而增强网络对微小判别特征的感知.DFFM模块探索特征间的相关性,对不同尺寸的特征交互建模,以得到更加丰富的互补信息,从而产生更强的细粒度特征.结果 本文方法可以端到端地进行训练,不需要边界框和多阶段训练.在CUB-200-2011(Caltech-UCSD Birds-200-2011)、Stanford Dogs、Stanford Cars 以及 FGVC-Aircraft(fine-grained visual classification of aircraft)4个基准数据集上验证所提方法,准确率分别达到了 92.6％、94.5％、95.3％和93.5％.实验结果表明,本文方法的性能优于当前主流方法,并在多个数据集中表现出良好的性能.在消融研究中,验证了模型中各个模块的有效性.结论 本文方法具有显著性能,通过注意互补有效提升了特征的多样性,以此尽可能地获取丰富的判别特征,使分类的结果更加精准.

全面深入提升高校教师能力素养，是新时代推进科教兴国战略、培养符合社会发展要求的高素质人才的客观需要。青年教师是高校师资力量的重要组成部分，本文从新时代高校青年教师担当使命的能力需求出发，设计了能力素养指标框架，从而以此为支撑，提出了全面培养新时代高校青年教师能力素养的基本策略。

Person re-identification mainly uses computer vision technology to determine whether there are specific pedestrians in the image or video. Belong to cross-device retrieval images, due to the changing style of the device led to more difficult person re-identification. The current algorithms use multi-feature fusion methods such as posture detection to match, ignoring the impact of different perspectives, postures and backgrounds on features. This paper proposes a multi-scale feature method based on saliency model. Which uses the salient image extraction algorithm to better filter out the interference of complex background parts, and uses the feature weighting method to fuse the global features and local features to achieve more robust features. Experimental results on three datasets show that the proposed method is superior to the existing method.