Search

With the rapid development of deep learning, face forgery detection methods have also achieved remarkable progress. However, most methods suffer significant performance degradation on low-quality compressed face images. It is due to: (a) The image artifacts will be blurred in the process of image compression, resulting in the model learning insufficient artifact traces; (b) Low-quality images will introduce a lot of noise information, and minimizing the training error causes the model to absorb all correlations in the training dataset recklessly, leading to the over-fitting problem. To solve the above problems, we consider learning domain invariant representations to inscribe the correct relevance, i.e., artifacts, to improve the robustness of low-quality images. Specifically, we propose a novel face forgery detector, called DIFLD. The model has the following components: (1) a high-frequency invariant feature learning module(hf-IFLM), which effectively retrieves the blurred artifacts in low-quality compressed images; and (2) a high-dimensional feature distribution learning module(hd-FDLM), that guides the network to learn more about the consistent features of distribution. With the above two modules, the whole framework can learn more discriminative correct artifact features in an end-to-end manner. Through extensive experiments, we show that our proposed method is more robust to image quality variations, especially in low-quality images. Our proposed method achieves a 3.67% improvement over the state-of-the-art methods on the challenging dataset NeuralTextures.

Bone milling surgery is one of the commonly used methods for the treatment of orthopedic diseases in clinics, and the use of robots to assist in the completion of bone milling surgery can effectively improve the efficiency of the operation. However, since the addition of robots will diminish the surgeon’s ability to perceive the surgical environment, improving the robotic system’s ability to perceive the milling state is an important factor in improving surgical safety. Inspired by the human auditory perception mechanism, this article proposes a method to recognize the milling state (including the type, depth, and angle of milling bones) using milling acoustic signals combined with neural networks. In this study, a series of bone milling experiments were completed to construct the datasets. The correlation between acoustic signal features and the milling state is analyzed, and then 17-D features are extracted as inputs to the network model. A novel auditory attention model combining the attention mechanism with the CNN-LSTM network for bone milling state recognition has been proposed, and the effects of different network structures on recognition accuracy are discussed. The results indicate that the proposed classification network has a great recognition accuracy for milling states, which proves that the method is beneficial for improving the safety of automatic milling surgery for robots.

During bent bone grinding, the robot often needs to adjust its grinding depth and angle in real time to ensure surgical quality. Optical- or vision-based navigation methods are no longer applicable due to the interference caused by splashing cooling water in the operating field. Inspired by orthopedic surgeons, in this article, we present a tactile perception-based method to solve this problem. The effect of grinding parameters on the cutter's current and vibration is analyzed, and the current and vibration signals are further processed to predict depth and angle. Depth and angle estimation models were calibrated on homogeneous artificial bone. Grinding control experiments were carried out on femur models, and the statistical mean and standard deviation results of the measured depth show that the accuracy of depth control can be further improved by adding angle estimation and proper adjustment when depth is indirectly estimated based on sensing signals. The proposed approach relies solely on the estimation of the relative position and posture between the tool and the bone surface and is expected to improve the safety of automatic grinding with orthopedic robots.

This article aims to optimize robotic actions for bone milling during laminectomy. The layer-by-layer spinal lamina milling operation during robot-assisted laminectomy is analyzed in detail to summarize related control tasks. A milling sound signal processing method based on band filters and the two-order Prony algorithm is proposed to overcome the frequency disturbance of the surgical power device's rotation motion caused by bone milling to accurately extract harmonic amplitudes. A depth estimation model based on the total value of major harmonics is fitted to monitor the actual depth when the ball end is cutting into an easily displaced bone. A milling feed speed optimization method based on the sound signal feedback and a fuzzy controller is designed to balance the efficiency and safety of bone milling. Milling experiments are carried out on artificial bone blocks and fresh swine cervical vertebrae. The results show that the depth monitoring accuracy is about 0.08 mm and the feed speed can be optimized effectively to match the actual bone mineral density and axial depth of cut. The proposed method can improve the depth accuracy and output a safe and fastest feed speed.

To address the lack of force measurement in minimally invasive robotic surgery, this article develops the sensorized surgical instrument based on fiber Bragg grating (FBG) technology. The sensorized surgical instrument detects forces in two directions by affixing three FBGs at 120° intervals on the shaft of the surgical instrument. For the designed sensorized surgical instrument, first, finite-element simulation is used to explore the impact of the trocar on the strain at the adhesive points of FBGs in the operation state. Subsequently, the study theoretically analyzes how the fixed position of the instrument on the calibration platform affects the strain. Furthermore, finite-element simulation is used to compare strain differences between the calibration state and the operation state. Finally, the static calibration experiment establishes the decoupling matrix for the instrument and demonstrates that changing its fixed position on the calibration platform yields a maximum force measurement error of 0.92%. Additionally, the accuracy of sensorized surgical instrument force measurement is demonstrated by hanging a weight from the end of the instrument and its surgical efficacy is proven in tumor localization tasks.

Development of combined mass spectrometry ionization sources has enabled expansion of the application and scope of mass spectrometry. A novel hybrid ionization system combining vacuum ultraviolet (VUV) and atmospheric pressure chemical ionization (APCI) was constructed. Gaseous samples were self-aspirated into an ionization zone through a capillary by negative pressure, generated by high-speed airflow based on the Venturi effect. Compared with APCI mode alone, the signal-to-noise ratio (S/N) in APCI/VUV mode was increased by about 276-times. To increase the ionization efficiency further, correlated experimental conditions were optimized. Four types of volatile organic compounds (VOCs) were tested to evaluate the performance of the APCI/VUV ion source. Excellent linearity and limit of detection were achieved for compounds in mixed solutions. Quantitative analyses of four VOCs (toluene, cyclohexanone, styrene and ethylbenzene) using APCI/VUV-MS were done, and the relative standard deviations (RSDs) were 1.57%, 6.30%, 4.49% and 8.21%, respectively, indicating that the APCI/VUV ionization source had excellent reproducibility. Our results demonstrated that the developed method was promising for analyzing VOCs as well as being rapid, simple, and easy to operate.

Due to the limited operation field and the deformation of the bone being cut, experienced surgeons usually will capture the bone layer and cutting depth information by analyzing the cutting acoustic signal (CAS). However, surgeons lose this sensing ability in robotic operations because they are no longer near the surgical area. This article aims to make the robot possess an acoustic signal-based monitoring ability such as the surgeon to estimate the proportion of different bone tissues and control cutting depth well. A quality index is designed to choose the optimal acoustic harmonic amplitude by analyzing the recorded acoustic signal in the idling and bone-cutting stages. Two mappings are fit between the optimal amplitude and the cutting depth on cancellous and cortical bones in advance. When the cutting area is located at mixed bone layers, the cutting depth can be further calculated based on the pre-obtained mappings and the bone layer proportion identified by the pre-trained regression network. A controller based on depth error is used to track the desired depth, while adaptive selection rules based on the proportion of cortical bones are introduced to optimize controller parameters to adapt the bone layer. Validation experiments of the proposed method via automatic cutting on thin artificial bone plates and pig scapula samples test the depth control accuracy and scope of application. The proposed method can improve the operation accuracy and safety of robotic bone milling.

In recent years, YOLOv5 networks have become a research focus in many fields because they are capable of outperforming state-of-the-art (SOTA) approaches in different computer vision tasks. Nevertheless, there is still room for improvement in YOLOv5 in terms of target tracking. We modified YOLOv5 according to the anchor-free paradigm to be on par with other state-of-the-art tracking paradigms and modified the network backbone to design an efficient module, thus proposing the RetinaYOLO detector, which, after combining state-of-the-art tracking algorithms, achieves state-of-the-art performance: we call it RetinaMOT. To the best of our knowledge, RetinaMOT is the first such approach. The anchor-free paradigm SOTA method for the YOLOv5 architecture and RetinaYOLO outperforms all lightweight YOLO architecture methods on the MS COCO dataset. In this paper, we show the details of the RetinaYOLO backbone, embedding Kalman filtering and the Hungarian algorithm into the network, with one framework used to accomplish two tasks. Our RetinaMOT shows that MOTA metrics reach 74.8, 74.1, and 66.8 on MOT Challenge MOT16, 17, and 20 test datasets, and our method is at the top of the list when compared with state-of-the-art methods.

Bone milling is commonly performed in many surgical interventions, but it often suffers from unpredictable deflection of the bone and potential damage to critical anatomical structures. In this study, a method is proposed to control the depth of cut based on the milling vibration monitored in real time. An accelerometer is mounted on the surgical power tool to record its vibration. In order to extract characteristic information from the acceleration signal, a quality factor is defined to select the optimal harmonic in the signal. A mapping is then established between the amplitude of the optimal harmonic and the depth of cut measured by a displacement sensor. During automatic milling in a single type of material, the depth of cut is estimated by inputting the current harmonic amplitude into the mapping, and the robot's position is adjusted based on the deviation between the desired and the estimated depth of cut. When the cutting area is located between the cancellous bone and the cortical bone, the depth of cut is calculated based on the mapping for the cancellous bone, the mapping for the cortical bone, and the percentage of these two tissues. Thus, automatic milling is also available in this area. Validation of the proposed method via automatic milling experiments shows effectiveness and stableness of the control algorithm, and the safety and efficiency are ensured.

To clarify the synergistic effect of well/over-tempered hierarchical martensitic structure and carbides (M23C6and MX) on austenite growth, the austenite growth during step-heat austenitization for a selected 9Cr steel was investigated by dilatometry, multi-scale characterization and cellular automaton simulation. The austenite grain inherits the geometrical characteristics of the hierarchical martensite substructure to a significant extent through covariant and free growth. In particular, the initially heterogeneous grains with non-uniform martensitic substructure can easily produce mixed and heterogeneous grains. Step isothermal heating under AC1can promote the evolution of martensitic structure to recrystallized ferrite and recovered equiaxed substructure, associated with coarsening of M23C6and precipitation of MC. Apart from the recrystallized ferrite promoting the growth of equiaxed austenite grains, the recovered approximate equiaxed substructure decorated with coarse M23C6and a large amount of MX can promote the subsequent development of fine spherical austenite grains by reducing the abnormal austenite growth and increasing the resistance to interfacial migration. The synergistic effect of the well/over tempered martensitic hierarchical structure and the carbides (M23C6, MX) is critical to the final austenite grain distribution. A window of critical temperature values for step heating austenitization to generate homogeneous austenite grain growth was established by exploiting the inheritance of the equiaxed substructure and reasonably reducing abnormal growth.

Degradation modeling-based remaining useful life (RUL) predicting has become a significant basis of prognotics and system health management (PHM) and has drawn much favor of both scholars and engineers in the field of reliability. Based on a purely time-dependent diffusion process, this article presents a nonlinear degradation model with a special focus on the proportional relationship between the age-dependent expectation and variance of the degradation processes of the concerned devices. Exact probability density function (pdf) and cumulative distribution function (cdf) of lifetime and RUL in explicit forms are derived under the concept of first hitting time (FHT), incorporating the item-to-item variability. A framework of maximizing the likelihood function has been proposed to estimate the unknown parameters utilizing condition monitoring (CM) data of the devices. Case studies of fatigue crack length data of alloy and capacity data of electrolytic capacitors have been presented to illustrate the proposed prognostics model. The results show that the proposed degradation model can be well fitted to the nonlinear degradation data and provide an accurate prediction of RUL, and thus demonstrate the effectiveness of the proposed model.

Various deep learning methods have been proposed to segment breast lesions from ultrasound images. However, similar intensity distributions, variable tumor morphologies and blurred boundaries present challenges for breast lesions segmentation, especially for malignant tumors with irregular shapes. Considering the complexity of ultrasound images, we develop an adaptive attention U-net (AAU-net) to segment breast lesions automatically and stably from ultrasound images. Specifically, we introduce a hybrid adaptive attention module (HAAM), which mainly consists of a channel self-attention block and a spatial self-attention block, to replace the traditional convolution operation. Compared with the conventional convolution operation, the design of the hybrid adaptive attention module can help us capture more features under different receptive fields. Different from existing attention mechanisms, the HAAM module can guide the network to adaptively select more robust representation in channel and space dimensions to cope with more complex breast lesions segmentation. Extensive experiments with several state-of-the-art deep learning segmentation methods on three public breast ultrasound datasets show that our method has better performance on breast lesions segmentation. Furthermore, robustness analysis and external experiments demonstrate that our proposed AAU-net has better generalization performance in the breast lesion segmentation. Moreover, the HAAM module can be flexibly applied to existing network frameworks. The source code is available on https://github.com/CGPxy/AAU-net.

Previous studies have explored effects of the conscious processing of emotional information on the interaction between emotion and executive control. However, the conscious processing of emotional information introduces “defensive responses,” which might limit the emotional effect. To avoid the limitations, this study aimed to explore the effect of automatic emotional processing on response inhibition among heroin abstainers. Individuals experiencing heroin withdrawal symptoms are not entirely exempt from the influence of drugs, and it is challenging to find heroin addicts in practice. Therefore, similar to previous studies, we recruited heroin abstainers in this study. In Experiment 1, for the heroin abstainers group, 30 participants meeting the inclusion criteria were recruited from a compulsory isolated detoxification center, whereas 30 age‐ and education‐matched participants were recruited from the community as the control group. These participants completed the Go/No‐Go task by supraliminal emotional priming. In Experiment 2, the heroin abstainers group comprised 33 eligible participants and control group comprised 33 eligible participants. These participants completed the Go/No‐Go task by subliminal emotional priming. We compared the participants’ rates of commission errors in the Go/No‐Go tasks. The rate of commission errors was lowest upon the activation of positive emotion. Moreover, under subliminal emotional priming, the rate of commission errors for the heroin abstainers group was significantly lower than that for the control group. However, regarding supraliminal emotional priming, there was no significant difference between the rates of commission errors for the two groups. These results suggest automatic positive emotional processing can enhance response inhibition among heroin abstainers. Moreover, under subliminal emotional priming, heroin abstainers had more enhanced response inhibition than the control groups. This study plays a significant role in understanding the effect of automatic emotional processing on response inhibition among drug addicts. This study also provides a theoretical basis for ensuring enhanced response inhibition among drug addicts.

Hydrogen atoms and SiClx(x= 1, 2, 3) radicals coexist during the hydrogenation of silicon tetrachloride (STC, SiCl4), an important process in the fabrication of industrial polysilicon. In this work, the mechanisms and kinetics of the reactions between H and SiClx(x= 1, 2, 3) were studied by theory. The structures and vibrational frequencies of reactants, products, intermediates, and transition states (TSs) were determined at the B2PLYP/may-cc-pVTZ level. The single-point energies of minima and saddle points were refined using the coupled-cluster single–double with triple perturbative (CCSD(T)) with the complete basis set extrapolation method. Some special treatments were designed to obtain reliable wave functions for unimolecular reactions without tight TSs by the density functional theory. Subsequently, Lennard-Jones (L-J) parameters between each intermediate (SiHClx) and bath gas (He) were obtained at the MP2/jul-cc-pVTZ level to derive reliable temperature- and pressure-dependent rate coefficients for unimolecular reactions according to the variational Rice–Ramsperger–Kassel–Marcus theory. For bimolecular reactions, rate coefficients were determined by the variational transition-state theory. The rate coefficients of barrierless reactions were derived based on the loose TSs with the maximum free energy. Finally, the master equation analysis was used to investigate the variation of the rate coefficients with pressure and temperature in the activated paths.

In response to the lack of objective evaluation criteria for interpreting the spatial scales of historical streets, as well as the problem of fragmented and complex textures, this research proposes an analysis method for assessing the spatial scaling effects using the Ping Ge type map. Drawing on the Ping Ge cartographic methods from the Qing Dynasty (1644–1911), it connects ancient measurement systems with objective evaluation criteria based on object relations and utilizes surveying maps from the Republic of China (1912–1949) to clarify temporal attributes of texture. The study employs a typical case study to demonstrate the feasibility of using the Ping Ge type map in interpreting historical streets. By creating the Ping Ge type map of the East‒West Street in Quanzhou Ancient City and utilizing form diagrams, it reveals patterns such as scale multiplications within respective plots, transformations in linear sequences of streets, and increases or decreases in plot series scales. The case analysis indicates that the Ping Ge type map effectively uncover the spatial scaling characteristics of historical street layouts in China and can transform them into design resources that preserve the inner order of cities, thereby promoting spatial scaling awareness in the planning and design of regional architectural clusters.

Three-dimensional (3D) porous hosts play pivotal roles in realizing dendrite-free lithium metal anodes (LMAs) owing to their high specific area. However, uneven local electric field and lack of lithiophilic sites on the reactive interface cause nonuniform lithium ion (Li+) deposition, leading to Li dendrite growth and parasitic reactions. These issues will inevitably incur short cycling life and severe safety hazards of lithium metal batteries. Herein, a 3D self-supporting host composed of hollow carbon nanofibers incorporated with silicon (Si) nanoparticles inside (Si-HCF) is constructed as Li metal host by a scalable coaxial electrospinning technique. During the first cycle, the Si nanoparticles is alloyed with Li+to form lithiophilic Li-Si alloy, contributing to deliver a low nucleation overpotential and homogeneous surface potential confirmed by Kelvin probe force microscopy, finally guiding Li+flux homogeneously throughout the fibrous mat. As expected, the Li metal can percolate through Si-HCF host reversibly without uncontrollable Li dendrites and the Si-HCF exhibits a stable cycle life of Li plating/stripping over 1400h. Additionally, the full cell combined with LiFePO4cathodes presents steady cycling over 400 cycles with a high average coulombic efficiency of 99.4%. This work provides new insights for the synthesis of 3D hosts with controllable nucleation sites to homogenize Li+deposition in LMAs.

Elastomers, known for their high stretchability and flexibility, are widely used in high-tech applications. However, traditional manufacturing methods for elastomeric part production have limitations. 3D printing, particularly fused deposition modeling (FDM), offers a promising alternative by allowing the fabrication of customized elastomers with desired shapes and properties. Conventional filament-based FDM techniques struggle to print elastomers. This article presents a novel approach for 3D printing polyolefin elastomer (POE) using a direct pellet printing technique. A customized pellet printer with a pneumatic pressure feeding system was used that eliminates filament buckling issues commonly associated with conventional filament-based 3D printing methods. The mechanical properties and microstructure of the printed parts were analyzed to evaluate the suitability of the technique for producing high-quality elastomeric components. SEM images indicated a high-quality and accurate printing method; however, there are micro-holes between the raster due to the high shrinkage rate of POE and increasing the nozzle temperature improves the print quality. The mechanical properties of the printed samples exhibited remarkable formability, with elongation reaching up to 1965%. It is also found that as the nozzle temperature increased, the strength, elongation, and bonding between layers improved significantly. This innovative 3D printing technique has the potential for various applications such as soft robotics and wearable electronics.

Melamine foam (MF), a promising development in light materials, finds application in construction, agriculture, aviation, transportation, electronic message, and other fields due to its excellent thermal insulation performance, sound absorption and noise reduction capabilities, safety and health benefits, as well as easy processing. However, its own shortcomings such as hardness, poor toughness, fragility, and slag removal greatly limit its application scope. In this review, a survey of the literature from two aspects of toughening of melamine resin and regulation of MF pore structure are reviewed to explore the research progress of toughening modification of MF. The principle, merit, and demerit of different modification methods are analyzed. In addition, owing to the extensive literature available, this article also summarizes the representative achievements of the nanotechnology modification of MF derivatives (carbon foam and carbon aerogel). Eventually, based on an assessment of current application status for both MF and its derivatives while considering existing challenges in their modification processes using nanotechnology approaches, we discuss future prospects for their application.

Direct analysis of highly reactive volatile species such as the aliphatic aldehydes as vital biomarkers remains a great challenge due to difficulties in the sample pretreatment. To address such a challenge, we herein report the development of a novel double-region atmospheric pressure chemical ionization mass spectrometry (DRAPCI-MS) method. The DRAPCI source implements a separated structural design that uses a focus electrode to divide the discharge and ionization region to reduce sample fragmentation in the ionization process. Counterflow introduction (CFI) configuration was adopted in the DRAPCI source to reduce background noise, while ion transmission efficiency was optimized through simulating the voltage of the focus electrode and the ion trajectory of the ion source. The limits of detection (LODs) of four carbonyl compounds cyclohexanone, hexanal, heptanal, and octanal by DRAPCI-MS were between 0.1 and 3 μg·m–3, approximately two to eight times lower than those by atmospheric pressure chemical ionization mass spectrometry. Additionally, the DRAPCI-MS method carried out effective in situ analyses of the volatile components in expired milk and the exhaled breath of smokers, demonstrating the DRAPCI-MS as a practical tool to analyze complex mixtures. The DRAPCI-MS method provides a rapid, sensitive, and high-throughput technique in the real-time analysis of gaseous small-molecule compounds.

The disposition and metabolism of nicotine in the brain is an important determinant of its exposure. We have developed a novel method for the dynamic determination of nicotine and its metabolites in rat brain and blood by simultaneous microdialysis sampling, stable-isotope labeling, and ultra high-performance liquid chromatography–high-resolution mass spectrometry (UHPLC–HRMS) assaying. Microdialysis probes were inserted into both the right striatum and jugular vein of Sprague–Dawley rats. The collections of dialystes after nicotine intraperitoneal injection were analyzed by the optimized UHPLC–HRMS. Nicotine-pyridyl-d4was used as a metabolic tracer, and several stably labeled isotopes were applied to calibrate the in vivo recoveries of retrodialysis. The quadrupole-Orbitrap HRMS provided reliable characterization of the nicotine derivatives with less than 3.5 ppm mass measurement accuracy. Good precision and accuracy were obtained for different analytes within the predefined limits of acceptability and the range of the standard curve. Nicotine and its 11 metabolites were identified in most microdialysis samples from the blood and brain tissue samples. Besides cotinine as the main metabolic product of nicotine, trans-3′-hydroxy-cotinine, nicotine-N-oxide, and norcotinine were proven to be the second most abundant metabolites. The other seven nicotine products, including 4-oxo-4-(3-pyridyl)-butanoic acid, 4-hydroxy-4-(3-pyridyl)-butanoic acid, cotinine-N-oxide, nicotine-N-glucuronide, cotinine-N-glucuronide, and trans-3′- hydroxy-cotinine-O-glucuronide, which have not been determined previously in animal brain, were present in minor amounts. The pharmacokinetic profile of nicotine metabolites indicated that the metabolic characteristic of nicotine in the brain was relatively different from that in the blood. The present work would provide comprehensive evidence for clarifying the differences between nicotine metabolism in the brain and peripheral system.

A novel method was established for rapid detection and confirmation of flavouring substances in flavoured pipe tobacco by a light-radiation thermal desorption-assisted low temperature plasma ionization source coupled with a miniature ion trap mass spectrometer (TD-LTP-ITMS). To improve the sensitivity of flavouring substances with low volatility, the heated light from a halogen lamp was applied to improve the partial pressure in the atmosphere. Four common flavours, vanillin, ethyl vanillin, tartaric acid and citric acid, were selected to demonstrate the signal intensities achieved with light-radiation heating. The experimental results prove that the signals were enhanced by more than 2.1 fold compared with those obtained at room temperature. Quantitative analysis was also performed, and the linear range was about two orders of magnitude. TD-LTP-ITMS was optimized to analyze each sample of the flavoured pipe tobacco and its extracted liquid within twenty seconds of heating. Meanwhile, vanillin and ethyl vanillin were confirmed by tandem mass spectrometry (MS/MS) in a complex matrix. The experimental results indicate that the developed method has the potential to rapidly determine the flavouring substances in complex matrices with high throughput.

With the popularity of online shopping and the rise of various brands, logo detection is gradually coming to the forefront of researchers' minds. However, accurately detecting multiscale, similar, diverse and shape-shifting logos poses a challenge for this technology. The Swin transformer has created a milestone as a high-performance deep learning method across modalities, domains and tasks in various tasks of computer vision. We improve the Swin transformer for better image feature extraction and enhance the robustness for its application to logo detection. In this paper, we propose the PMST as the backbone of logo detection, and design a bypass-parallelizable shift module and a miniature window tandem shift strategy to further enhance image feature fusion and transfer between windows. The PMST achieves a 79.2 box AP on the LogoData dataset, surpassing most detection models. According to existing known methods, in logo detection, the PMST is the first backbone to put transformer to the detection. The results achieved state-of-the-art on the FlickrLogos-32 and FoodLogoDet-1500 datasets. It also achieves excellent efficiency on the ImageNet-1K, OpenBrands-80 and MS-COCO datasets. The code is available at https://github.com/blowhen/PMST. [ABSTRACT FROM AUTHOR]

Multi-target tracking is a complex problem in computer vision, involving occlusion, target trajectory disappearance, etc. In response to these questions, we first propose a trajectory coordinate bi-directional sensing attention module (BasicCaModule). This module enables the network to be more focused on essential areas. Secondly, we propose the Feature Confusion Edge Enhancement (FCEE) module, which enhances the confusion region of adjacent feature edges. After detecting the network output feature map, we propose the adaptive feature separation header (FDH) module. This module can activate essential features adaptively to reduce the number of target identity switches. Finally, appropriate hyperparameters are set and optimized for the ByteTrack tracking algorithm, resolving fragment tracking loss and long memory matching problems. After a large amount of experimental data, our approach is highly competitive with the existing state-of-the-art algorithms. Moreover, it can also keep online tracking while maintaining high tracking accuracy. The code is publicly available at: https://github.com/CccJ23/Spatio-temporal-trajectory-Tracking [ABSTRACT FROM AUTHOR]

In this paper, the dynamic response of fiber-metal laminate (FML) double-layer sandwich plates under blast loading is studied through theoretical analysis and finite element (FE) calculation. The membrane mode solution is obtained for the dynamic response of the clamped FML double-layer sandwich plate under blast loading, and the so-called ‘bounds’ are obtained through circumscribing and inscribing lines of the exact yield locus. The FE model is established, and the analytical model is proved by the FE method. The influences of geometrical parameters and material properties on the dynamic response of FML double-layer sandwich plates are researched based on the analytical model. Finally, it is found that FML double-layer sandwich plates have better anti-explosion performance with equal mass by comparing with the metal sandwich plates.

Under low-velocity impact, the dynamic response of fully clamped foam-filled sandwich beam with circular shape core is studied theoretically and numerically in this paper. Based on the yield condition of the foam-filled sandwich structure, the theoretical model of the dynamic response of the foam-filled sandwich beam with circular shape core under low-velocity impact is established, considering the foam effect and the interaction between stretching and bending moment. The ‘bound’ and analytical solution for the dynamic response of the foam-filled sandwich beam with circular shape core are derived. To verify the correctness of the theoretical solution, the finite element analysis is carried out. It is found that the difference between the two is very small, which can be considered that they are consistent. Finally, the effects of geometric parameters, material properties and the impact location about dynamic response of the foam-filled sandwich beams with circular shape core are discussed. It is shown that increasing the strength and thickness of face-sheets, the strength of foam, the strength and thickness of circular tube result in the increase in the impact force for the given deflection. When the loading location is farther from the midspan the sandwich beam filled with foam, the impact force increases, the maximum deflection decreases for the given impact energy.

Experiments on splitting and curling collapse of square sandwich metal tubes with aluminum foam core under axial low-velocity impact are conducted in this paper. Experimental details are described including geometric and test parameters of specimens. Deformation modes and load–displacement curves are obtained through experiment. Analytical solution of quasi-static splitting and curling steady load can be used to predict steady load of square sandwich tube under axial low-velocity impact based on the similarity of quasi-static and low-velocity impact deformation modes.

Once in the hands of end users, such durable equipment as spacecraft, aircraft, ships, automobiles, computers, etc. are in a state of debugging, working or storage. In either state, availability, reliability and super-efficiency are the ultimate goals, which have been achieved through constant monitoring as well as regular, preventive, routine and corrective maintenance. Although some advanced instruments can visualize certain invisible malfunctioning phenomena into visible ones, deeply hidden troubles cannot be found unless monitoring and testing data are addressed using tools that process the data statistically, analytically and mathematically. Some state-of-the-art trouble-shooting and life-predicting techniques and approaches are introduced in this paper.

The plastic behaviour (PB) of fully clamped foam-filled X-shaped core sandwich beams (SBs) is researched analytically and numerically. Considering the strengths of X-shaped core and foam, the yield criterion of foam-filled X-shaped core sandwich structure is presented. Taking into account the interaction between traction and bending moment caused by large deflection, the analytical solution of foam-filled X-shaped core SBs loading laterally under a punch is derived. The numerical results calculated by Abaqus/Standard software are in great accordance with the analysis. The influences of geometrical parameters and foam strength about the load-bearing capability and absorbed deformation energy capability of the foam-filled X-shaped core SB are clearly discussed. The results show that the axial stretching caused by large displacement has a significant impact on load-bearing capability and absorbed deformation energy of foam-filled X-shaped core SBs. This analysis approach could be used to predict the PB of foam-filled X-shaped core SB.

The realistic degradation process for the engineering equipment is generally stochastic and complicated owing to the uncertain operational condition and multiple functional loading, exhibiting the absolute nonlinear distinction. Such a nonlinear degradation process is widely modeled as a generalized diffusion process. When utilizing the generalized diffusion process-based model, certain model parameters are considered as the random variables to characterize the unit-to-unit discrepancies. Hence, the estimation of these kinds of parameters usually resorts to the Bayesian method. However, owing to the complex pattern of the model parameters in the generalized diffusion process, computing the Bayesian updated parameters requires plenty of repeated calculation and integration operations once the new degradation monitoring information is available. This will inevitably lower the computing efficiency and real-time performance. Toward this end, this paper presents an adaptive prognostic method based on the generalized diffusion process to determine the remaining useful life (RUL) of degraded equipment. First, a generalized diffusion process-based degradation modeling framework is constructed to describe the health performance of stochastic degraded equipment under complex conditions. Then, we utilize the maximum likelihood estimation (MLE) method to estimate the initial model parameters by analyzing the historical degradation information. Furthermore, a sequential Bayesian method is proposed to recursively update the stochastic model parameters in the generalized diffusion process for particular equipment in service. Unlike the existing studies utilizing the Bayesian method, the primary contrast in the presented method lies in that there is no need to implement the calculation process with complicated integration repeatedly utilizing the whole degradation information obtained before the current time. Particularly, the current measured information is incorporated into the estimates of the stochastic parameters in the previous time to determine the corresponding posterior estimates at the current time. This can avoid repeated calculation and raise the efficiency to a certain extent. Thereafter, the RUL distribution is updated adaptively by incorporating the acquired posterior estimates. Finally, we provide two practical case studies associated with the gyroscope and 2017-T4 aluminum alloy to demonstrate the efficiency and advantage of the proposed sequential Bayesian method. The experimental results exhibit that the proposed method can increase the RUL prediction accuracy compared with the existing methods in the literature.

In this paper, the convex distortion and transient distortion characteristics of TIG welding for a titanium alloy with finite dimensional thin plate were investigated experimentally and numerically. A three dimensional thermal-mechanical finite element method (FEM) was adopted to predict the characteristics of Transient distortion by taking the initial shape, gravity loading and restriction conditions into account. The simulation results were validated by the experimental results. The results show that the numerical simulations could provide accurate prediction on the Transient temperature and distortion processing. The temperature gradient through thickness, the dimension of plate and the clamping were main factors governing the distortion during welding.

Recent developments in prognostic and health management have been targeted at utilizing the observed degradation signals to estimate residual life distributions. Current degradation models mainly focus on a population of “identical” devices or an individual device with population information, not a single component in the absence of prior degradation knowledge. However, the fast development of science and technology provides us with many kinds of new systems, and we just have the real-time monitoring information to analyze the reliability for them. The fusion algorithm presented herein addresses this challenge by combining the excellent modeling ability of Bayesian updating method for the multilevel data and the prominent estimation ability of ECM algorithm for incomplete data. Residual life distributions and posterior distributions are first calculated through the Bayesian updating method based on random initial a priori distributions. Then the a priori distributions are revised and improved for future predictions by the ECM algorithm. Once a new signal is observed, we can reuse the fusion algorithm to improve the accuracy of residual life distributions. The applicability of this fusion algorithm is validated by a set of simulation experiments.

A short proof of Nash-Williams' Theorem on the arboricity of a graph is given.

Ultrasound imaging has been widely used in the screening of kidney diseases. The segmentation of the kidney in ultrasound images can aid in the diagnosis of clinical diseases. However, segmenting the kidney from ultrasound images is challenging, due to the effects of heterostructure, image quality, and other factors. In this paper, a multi-branch and deep supervision network (MBDSNet) is proposed to segment the kidney from ultrasound images. Specifically, the architecture consists of two parts: multi-branch encoding network and refined decoding network. The multi-branch encoding network consists of a multi-scale feature input pyramid and three branch encoders. The multi-branch encoding network can reduce the sensitivity of the network to the input data through the multi-scale feature inputs pyramid, and utilize the information exchange between the three branch encoders can improve the generalization ability of the network. In the refined decoding network, the introduction of the side-out deep supervision module and the kidney boundary detection module can further guide the network to effectively segment the kidney with complete contour. Objective comparisons with several state-of-the-art segmentation methods on six commonly used metrics are performed. The method proposed in this paper has achieved the best results in terms of evaluation indicators and segmentation results. [ABSTRACT FROM AUTHOR]