An improved model based on you only look once version 8(YOLOv8)is proposed to solve the problem of low detection accuracy due to the diversity of object sizes in optical remote sensing images.Firstly,the feature pyram...An improved model based on you only look once version 8(YOLOv8)is proposed to solve the problem of low detection accuracy due to the diversity of object sizes in optical remote sensing images.Firstly,the feature pyramid network(FPN)structure of the original YOLOv8 mode is replaced by the generalized-FPN(GFPN)structure in GiraffeDet to realize the"cross-layer"and"cross-scale"adaptive feature fusion,to enrich the semantic information and spatial information on the feature map to improve the target detection ability of the model.Secondly,a pyramid-pool module of multi atrous spatial pyramid pooling(MASPP)is designed by using the idea of atrous convolution and feature pyramid structure to extract multi-scale features,so as to improve the processing ability of the model for multi-scale objects.The experimental results show that the detection accuracy of the improved YOLOv8 model on DIOR dataset is 92%and mean average precision(mAP)is 87.9%,respectively 3.5%and 1.7%higher than those of the original model.It is proved the detection and classification ability of the proposed model on multi-dimensional optical remote sensing target has been improved.展开更多
We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hie...We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hierarchical efficient multi-scale attention(H-EMA) module is designed for lightweight feature extraction, achieving outstanding performance at a relatively low cost. Secondly, an improved EfficientNetV2 block is used to integrate information from different scales better and enhance inter-layer message passing. Furthermore, introducing the convolutional block attention module(CBAM) enhances the model's perception of critical features, optimizing its generalization ability. Lastly, Focal Loss is introduced to adjust the weights of complex samples to address the issue of imbalanced categories in the dataset, further improving the model's performance. The model achieved 96.11% accuracy on the intertidal marine organism dataset of Nanji Islands and 84.78% accuracy on the CIFAR-100 dataset, demonstrating its strong generalization ability to meet the demands of oceanic biological image classification.展开更多
Multi-label image classification is a challenging task due to the diverse sizes and complex backgrounds of objects in images.Obtaining class-specific precise representations at different scales is a key aspect of feat...Multi-label image classification is a challenging task due to the diverse sizes and complex backgrounds of objects in images.Obtaining class-specific precise representations at different scales is a key aspect of feature representation.However,existing methods often rely on the single-scale deep feature,neglecting shallow and deeper layer features,which poses challenges when predicting objects of varying scales within the same image.Although some studies have explored multi-scale features,they rarely address the flow of information between scales or efficiently obtain class-specific precise representations for features at different scales.To address these issues,we propose a two-stage,three-branch Transformer-based framework.The first stage incorporates multi-scale image feature extraction and hierarchical scale attention.This design enables the model to consider objects at various scales while enhancing the flow of information across different feature scales,improving the model’s generalization to diverse object scales.The second stage includes a global feature enhancement module and a region selection module.The global feature enhancement module strengthens interconnections between different image regions,mitigating the issue of incomplete represen-tations,while the region selection module models the cross-modal relationships between image features and labels.Together,these components enable the efficient acquisition of class-specific precise feature representations.Extensive experiments on public datasets,including COCO2014,VOC2007,and VOC2012,demonstrate the effectiveness of our proposed method.Our approach achieves consistent performance gains of 0.3%,0.4%,and 0.2%over state-of-the-art methods on the three datasets,respectively.These results validate the reliability and superiority of our approach for multi-label image classification.展开更多
Because the physiological characteristics and melanin regulation mechanism of zebrafish are highly similar with those of humans,it is of high reference value to use zebrafish model in the evaluation of cosmetic whiten...Because the physiological characteristics and melanin regulation mechanism of zebrafish are highly similar with those of humans,it is of high reference value to use zebrafish model in the evaluation of cosmetic whitening efficacy.In this study,zebrafish embryos are used as biological models to evaluate the whitening efficacy of six kinds of cosmetics raw materials,such as antioxidant,preservative and essence,and the formula of facial cleanser and facial mask products,and the limitations of the zebrafish melanin production grayscale detection method in practical application are discussed.The results show that the selection of different types of components can also reduce the production of melanin and show whitening effect.It can be seen that the gray scale method of melanin production in zebrafish is suitable for the evaluation of the efficacy of raw materials.In practical application,due to the complexity of the formula,the toxic effects of different types of ingredients may interfere with the melanin generation of zebrafish,affecting the judgment and evaluation of whitening efficacy.For the detection of whitening efficacy of products,a comprehensive evaluation system should be built together with other methods to accurately evaluate the whitening efficacy.展开更多
To improve image quality under low illumination conditions,a novel low-light image enhancement method is proposed in this paper based on multi-illumination estimation and multi-scale fusion(MIMS).Firstly,the illuminat...To improve image quality under low illumination conditions,a novel low-light image enhancement method is proposed in this paper based on multi-illumination estimation and multi-scale fusion(MIMS).Firstly,the illumination is processed by contrast-limited adaptive histogram equalization(CLAHE),adaptive complementary gamma function(ACG),and adaptive detail preserving S-curve(ADPS),respectively,to obtain three components.Then,the fusion-relevant features,exposure,and color contrast are selected as the weight maps.Subsequently,these components and weight maps are fused through multi-scale to generate enhanced illumination.Finally,the enhanced images are obtained by multiplying the enhanced illumination and reflectance.Compared with existing approaches,this proposed method achieves an average increase of 0.81%and 2.89%in the structural similarity index measurement(SSIM)and peak signal-to-noise ratio(PSNR),and a decrease of 6.17%and 32.61%in the natural image quality evaluator(NIQE)and gradient magnitude similarity deviation(GMSD),respectively.展开更多
The joint roughness coefficient(JRC) is one of the key parameters for evaluating the shear strength of rock joints.Because of the scale effect in the JRC,reliable JRC values are of great importance for most rock engin...The joint roughness coefficient(JRC) is one of the key parameters for evaluating the shear strength of rock joints.Because of the scale effect in the JRC,reliable JRC values are of great importance for most rock engineering projects.During the collection process of JRC samples,the redundancy or insufficiency of representative rock joint surface topography(RJST) information in serial length JRC samples is the essential reason that affects the reliability of the scale effect results.Therefore,this paper proposes an adaptive sampling method,in which we use the entropy consistency measure Q(a) to evaluate the consistency of the joint morphology information contained in adjacent JRC samples.Then the sampling interval is automatically adjusted according to the threshold Q(at) of the entropy consistency measure to ensure that the degree of change of RJST information between JRC samples is the same,and ultimately makes the representative RJST information in the collected JRC samples more balanced.The application results of actual cases show that the proposed method can obtain the scale effect in the JRC efficiently and reliably.展开更多
This study reports the analytical solution for a generalized rotational pendulum system with gallows and periodic excited forces.The multiple scales method(MSM)is applied to solve the proposed problem.Several types of...This study reports the analytical solution for a generalized rotational pendulum system with gallows and periodic excited forces.The multiple scales method(MSM)is applied to solve the proposed problem.Several types of rotational pendulum oscillators are studied and talked about in detail.These include the forced damped rotating pendulum oscillator with gallows,the damped standard simple pendulum oscillator,and the damped rotating pendulum oscillator without gallows.The MSM first-order approximations for all the cases mentioned are derived in detail.The obtained results are illustrated with concrete numerical examples.The first-order MSM approximations are compared to the fourth-order Runge-Kutta(RK4)numerical approximations.Additionally,the maximum error is estimated for the first-order approximations obtained through the MSM,compared to the numerical approximations obtained by the RK4 method.Furthermore,we conducted a comparative analysis of the outcomes obtained by the used method(MSM)and He-MSM to ascertain their respective levels of precision.The proposed method can be applied to analyze many strong nonlinear oscillatory equations.展开更多
In the analysis of functionally graded materials （FGMs）, the uncoupled approach is used broadly, which is based on homogenized material property and ignores the effect Of local micro-structural interaction. The high...In the analysis of functionally graded materials （FGMs）, the uncoupled approach is used broadly, which is based on homogenized material property and ignores the effect Of local micro-structural interaction. The higher-order theory for FGMs （HOTFGM） is a coupled approach that explicitly takes the effect of micro-structural gradation and the local interaction of the spatially variable inclusion phase into account. Based on the HOTFGM, this article presents a quadrilateral element-based method for the calculation of multi-scale temperature field （QTF）. In this method, the discrete cells are quadrilateral including rectangular while the surface-averaged quantities are the primary variables which replace the coefficients employed in the temperature function. In contrast with the HOTFGM, this method improves the efficiency, eliminates the restriction of being rectangular cells and expands the solution scale. The presented results illustrate the efficiency of the QTF and its advantages in analyzing FGMs.展开更多
The natural frequencies of an axially moving beam were determined by using the method of multiple scales. The method of second-order multiple scales could be directly applied to the governing equation if the axial mot...The natural frequencies of an axially moving beam were determined by using the method of multiple scales. The method of second-order multiple scales could be directly applied to the governing equation if the axial motion of the beam is assumed to be small. It can be concluded that the natural frequencies affected by the axial motion are proportional to the square of the velocity of the axially moving beam. The results obtained by the perturbation method were compared with those given with a numerical method and the comparison shows the correctness of the multiple-scale method if the velocity is rather small.展开更多
The greatest challenges of rigorously modeling coupled hydro-mechanical(HM)processes in fractured geological media at different scales are associated with computational geometry.These challenges include dynamic sheari...The greatest challenges of rigorously modeling coupled hydro-mechanical(HM)processes in fractured geological media at different scales are associated with computational geometry.These challenges include dynamic shearing and opening of intersecting fractures at discrete fracture scales as a result of coupled processes,and contact alteration along rough fracture surfaces that triggers structural and physical changes of fractures at micro-asperity scale.In this paper,these challenges are tackled by developing a comprehensive modeling approach for coupled processes in fractured geological media based on numerical manifold method(NMM)at multiple scales.Based on their distinct geometric features,fractures are categorized into three different scales:dominant fracture,discrete fracture,and discontinuum asperity scales.Here the scale is relative,that of the fracture relative to that of the research interest or domain.Different geometric representations of fractures at different scales are used,and different governing equations and constitutive relationships are applied.For dominant fractures,a finite thickness zone model is developed to treat a fracture as a porous nonlinear domain.Nonlinear fracture mechanical behavior is accurately modeled with an implicit approach based on strain energy.For discrete fractures,a zero-dimensional model was developed for analyzing fluid flow and mechanics in fractures that are geometrically treated as boundaries of the rock matrix.With the zero-dimensional model,these fractures can be modeled with arbitrary orientations and intersections.They can be fluid conduits or seals,and can be open,bonded or sliding.For the discontinuum asperity scale,the geometry of rough fracture surfaces is explicitly represented and contacts involving dynamic alteration of contacts among asperities are rigorously calculated.Using this approach,fracture alteration caused by deformation,re-arrangement and sliding of rough surfaces can be captured.Our comprehensive model is able to handle the computational challenges with accurate representation of intersections and shearing of fractures at the discrete fracture scale and rigorously treats contacts along rough fracture surfaces at the discontinuum asperity scale.With future development of three-dimensional(3D)geometric representation of discrete fracture networks in porous rock and contacts among multi-body systems,this model is promising as a basis of 3D fully coupled analysis of fractures at multiple scales,for advancing understanding and optimizing energy recovery and storage in fractured geological media.展开更多
There are abundant igneous gas reservoirs in the South China Sea with significant value of research,and lithology classification,mineral analysis and porosity inversion are important links in reservoir evaluation.Howe...There are abundant igneous gas reservoirs in the South China Sea with significant value of research,and lithology classification,mineral analysis and porosity inversion are important links in reservoir evaluation.However,affected by the diverse lithology,complicated mineral and widespread alteration,conventional logging lithology classification and mineral inversion become considerably difficult.At the same time,owing to the limitation of the wireline log response equation,the quantity and accuracy of minerals can hardly meet the exploration requirements of igneous formations.To overcome those issues,this study takes the South China Sea as an example,and combines multi-scale data such as micro rock slices,petrophysical experiments,wireline log and element cutting log to establish a set of joint inversion methods for minerals and porosity of altered igneous rocks.Specifically,we define the lithology and mineral characteristics through core slices and mineral data,and establish an igneous multi-mineral volumetric model.Then we determine element cutting log correction method based on core element data,and combine wireline log and corrected element cutting log to perform the lithology classification and joint inversion of minerals and porosity.However,it is always difficult to determine the elemental eigenvalues of different minerals in inversion.This paper uses multiple linear regression methods to solve this problem.Finally,an integrated inversion technique for altered igneous formations was developed.The results show that the corrected element cutting log are in good agreement with the core element data,and the mineral and porosity results obtained from the joint inversion based on the wireline log and corrected element cutting log are also in good agreement with the core data from X-ray diffraction.The results demonstrate that the inversion technique is applicable and this study provides a new direction for the mineral inversion research of altered igneous formations.展开更多
The solution to heat transfer problems in two-dimensional heterogeneous media is attended based on the scaled boundary finite element method(SBFEM)coupled with equilibrated basis functions(EqBFs).The SBFEM reduces the...The solution to heat transfer problems in two-dimensional heterogeneous media is attended based on the scaled boundary finite element method(SBFEM)coupled with equilibrated basis functions(EqBFs).The SBFEM reduces the model order by scaling the boundary solution onto the inner element.To this end,tri-lateral elements are emanated from a scaling center,followed by the development of a semi-analytical solution along the radial direction and a finite element solution along the circumferential/boundary direction.The discretization is thus limited to the boundaries of the model,and the semi-analytical radial solution is found through the solution of an eigenvalue problem,which restricts the methods’applicability to heterogeneous media.In this research,we first extracted the SBFEM formulation considering the heterogeneity of the media.Then,we replaced the semi-analytical radial solution with the EqBFs and removed the eigenvalue solution step from the SBFEM.The varying coefficients of the partial differential equation(PDE)resulting from the heterogeneity of the media are replaced by a finite series in the radial and circumferential directions of the element.A weighted residual approach is applied to the radial equation.The equilibrated radial solution series is used in the new formulation of the SBFEM.展开更多
The present study proposes a sub-grid scale model for the one-dimensional Burgers turbulence based on the neuralnetwork and deep learning method.The filtered data of the direct numerical simulation is used to establis...The present study proposes a sub-grid scale model for the one-dimensional Burgers turbulence based on the neuralnetwork and deep learning method.The filtered data of the direct numerical simulation is used to establish thetraining data set,the validation data set,and the test data set.The artificial neural network(ANN)methodand Back Propagation method are employed to train parameters in the ANN.The developed ANN is applied toconstruct the sub-grid scale model for the large eddy simulation of the Burgers turbulence in the one-dimensionalspace.The proposed model well predicts the time correlation and the space correlation of the Burgers turbulence.展开更多
The construction of an ecological security pattern(ESP)is an effective approach to addressing ecological concerns in urban agglomerations.However,previous research has often overlooked the ecological characteristics a...The construction of an ecological security pattern(ESP)is an effective approach to addressing ecological concerns in urban agglomerations.However,previous research has often overlooked the ecological characteristics at different spatial scales,making it difficult to effectively tackle environmental issues.This study focuses on the Changsha–Zhuzhou–Xiangtan urban agglomeration(CZXUA)to construct an ESP from 2 scales:the urban agglomeration and the metropolitan area.First,ecological sources at different scales were identified and optimized using morphological spatial pattern analysis,ecosystem service supply assessment,landscape connectivity evaluation,and ecological demand intensity analysis.Next,spatial principal component analysis was employed to analyze resistance factors across different spatial scales,enabling the construction of a multi-scale landscape resistance surface.Finally,multi-scale ecological corridors were extracted using circuit theory,and the ecological corridors passing through the central urban area were optimized.As a result,176 ecological sources,337 ecological corridors,60 ecological pinch points,and 30 ecological barrier points were identified,forming the overall ESP for the CZXUA.Furthermore,an ecological protection and optimization pattern,described as“One River,Seven Waters,One Heart,and Eight Wedges”,was proposed based on the ESP.This pattern highlights pivotal areas for ecological protection and restoration within the national land space.This study provides a comprehensive example of constructing a multi-scale ecological corridor network and identifying critical areas for ecological protection and restoration.The findings can serve as a reference for special ecological protection planning in urban agglomerations.展开更多
In this investigation,some different approaches are implemented for analyzing a generalized forced damped complex Duffing oscillator,including the hybrid homotopy perturbation method(H-HPM),which is sometimes called t...In this investigation,some different approaches are implemented for analyzing a generalized forced damped complex Duffing oscillator,including the hybrid homotopy perturbation method(H-HPM),which is sometimes called the Krylov-Bogoliubov-Mitropolsky(KBM)method and the multiple scales method(MSM).All mentioned methods are applied to obtain some accurate and stable approximations to the proposed problem without decoupling the original problem.All obtained approximations are discussed graphically using different numerical values to the relevant parameters.Moreover,all obtained approximate solutions are compared with the 4thorder Runge-Kutta(RK4)numerical approximation.The maximum residual distance error(MRDE)is also estimated,in order to verify the high accuracy of the obtained analytic approximations.展开更多
The Balise Transmission Module(BTM)unit of the on-board train control system is a crucial component.Due to its unique installation position and complex environment,this unit has a higher fault rate within the on-board...The Balise Transmission Module(BTM)unit of the on-board train control system is a crucial component.Due to its unique installation position and complex environment,this unit has a higher fault rate within the on-board train control system.To conduct fault prediction for the BTM unit based on actual fault data,this study proposes a prediction method combining reliability statistics and machine learning,and achieves the fusion of prediction results from different dimensions through multi-method interactive validation.Firstly,a method for predicting equipment fault time targeting batch equipment is introduced.This method utilizes reliability statistics to construct a model of the remaining faultless operating time distribution considering uncertainty,thereby predicting the remaining faultless operating probability of the BTM unit.Secondly,considering the complexity of the BTM unit’s fault mechanism,the small sample size of fault cases,and the potential presence of multiple fault features in fault text records,an individual-oriented fault prediction method based on Bayesian-optimized Gradient Boosting Regression Tree(Bayes-GBRT)is proposed.This method achieves better prediction results compared to linear regression algorithms and random forest regression algorithms,with an average absolute error of only 0.224 years for predicting the fault time of this type of equipment.Finally,a multi-method interactive validation approach is proposed,enabling the fusion and validation of multi-dimensional results.The results indicate that the predicted fault time and the actual fault time conform to a log-normal distribution,and the parameter estimation results are basically consistent,verifying the accuracy and effectiveness of the prediction results.The above research findings can provide technical support for the maintenance and modification of BTM units,effectively reducing maintenance costs and ensuring the safe operation of high-speed railway,thus having practical engineering value for preventive maintenance.展开更多
Ceramic Matrix Composite （CMC） turbine guide vanes possess multi-scale stress and strain with inhomogeneity at the microscopic scale. Given that the macroscopic distribution cannot reflect the microscopic stress flu...Ceramic Matrix Composite （CMC） turbine guide vanes possess multi-scale stress and strain with inhomogeneity at the microscopic scale. Given that the macroscopic distribution cannot reflect the microscopic stress fluctuation, the macroscopic method fails to meet the requirements of stress and strain analysis of CMC turbine guide vanes. Furthermore, the complete thermodynamic properties of 2D woven SiC/SiC-CMC cannot be obtained through experimentation, Accordingly, a method to calculate the thermodynamic properties of CMC and analyze multi-scale stress and strain of the turbine guide vanes should be established. In this study, the multi-scale thermodynamic analysis is investigated. The thermodynamic properties of Chemical Vapor Infiltration （CVI） pro- cessed SiC/SiC-CMC are predicted by a Representative Volume Element （RVE） model with porosity, leading to the result that the relative error between the calculated in-plane tensile modulus and the experimental value is 4.2%. The macroscopic response of a guide vane under given conditions is predicted. The relative error between the predicted strain on the trailing edge and the experimental value is 9.7%. The calculation of the stress distribution of micro-scale RVE shows that the maximum value of microscopic stress, which is located in the interlayer matrix, is more than 1.5 times that of macroscopic stress in the same direction and the microscopic stress distribution of the interlayer matrix is related to the pore distribution of the composite.展开更多
A straightforward multi-scale boundary element method is proposed for global and local mechanical analysis of heterogeneous material.The method is more accurate and convenient than finite element based multi-scale met...A straightforward multi-scale boundary element method is proposed for global and local mechanical analysis of heterogeneous material.The method is more accurate and convenient than finite element based multi-scale method.The formulations of this method are derived by combining the homogenization approach and the fundamental equations of boundary element method.The solution gives the convenient formulations to compute global elastic constants and the local stress field.Finally,two numerical examples of porous material are presented to prove the accuracy and the efficiency of the proposed method.The results show that the method does not require the iteration to obtain the solution of the displacement in micro level.展开更多
Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware reso...Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.展开更多
基金supported by the National Natural Science Foundation of China(No.62241109)the Tianjin Science and Technology Commissioner Project(No.20YDTPJC01110)。
文摘An improved model based on you only look once version 8(YOLOv8)is proposed to solve the problem of low detection accuracy due to the diversity of object sizes in optical remote sensing images.Firstly,the feature pyramid network(FPN)structure of the original YOLOv8 mode is replaced by the generalized-FPN(GFPN)structure in GiraffeDet to realize the"cross-layer"and"cross-scale"adaptive feature fusion,to enrich the semantic information and spatial information on the feature map to improve the target detection ability of the model.Secondly,a pyramid-pool module of multi atrous spatial pyramid pooling(MASPP)is designed by using the idea of atrous convolution and feature pyramid structure to extract multi-scale features,so as to improve the processing ability of the model for multi-scale objects.The experimental results show that the detection accuracy of the improved YOLOv8 model on DIOR dataset is 92%and mean average precision(mAP)is 87.9%,respectively 3.5%and 1.7%higher than those of the original model.It is proved the detection and classification ability of the proposed model on multi-dimensional optical remote sensing target has been improved.
基金supported by the National Natural Science Foundation of China (Nos.61806107 and 61702135)。
文摘We propose a hierarchical multi-scale attention mechanism-based model in response to the low accuracy and inefficient manual classification of existing oceanic biological image classification methods. Firstly, the hierarchical efficient multi-scale attention(H-EMA) module is designed for lightweight feature extraction, achieving outstanding performance at a relatively low cost. Secondly, an improved EfficientNetV2 block is used to integrate information from different scales better and enhance inter-layer message passing. Furthermore, introducing the convolutional block attention module(CBAM) enhances the model's perception of critical features, optimizing its generalization ability. Lastly, Focal Loss is introduced to adjust the weights of complex samples to address the issue of imbalanced categories in the dataset, further improving the model's performance. The model achieved 96.11% accuracy on the intertidal marine organism dataset of Nanji Islands and 84.78% accuracy on the CIFAR-100 dataset, demonstrating its strong generalization ability to meet the demands of oceanic biological image classification.
基金supported by the National Natural Science Foundation of China(62302167,62477013)Natural Science Foundation of Shanghai(No.24ZR1456100)+1 种基金Science and Technology Commission of Shanghai Municipality(No.24DZ2305900)the Shanghai Municipal Special Fund for Promoting High-Quality Development of Industries(2211106).
文摘Multi-label image classification is a challenging task due to the diverse sizes and complex backgrounds of objects in images.Obtaining class-specific precise representations at different scales is a key aspect of feature representation.However,existing methods often rely on the single-scale deep feature,neglecting shallow and deeper layer features,which poses challenges when predicting objects of varying scales within the same image.Although some studies have explored multi-scale features,they rarely address the flow of information between scales or efficiently obtain class-specific precise representations for features at different scales.To address these issues,we propose a two-stage,three-branch Transformer-based framework.The first stage incorporates multi-scale image feature extraction and hierarchical scale attention.This design enables the model to consider objects at various scales while enhancing the flow of information across different feature scales,improving the model’s generalization to diverse object scales.The second stage includes a global feature enhancement module and a region selection module.The global feature enhancement module strengthens interconnections between different image regions,mitigating the issue of incomplete represen-tations,while the region selection module models the cross-modal relationships between image features and labels.Together,these components enable the efficient acquisition of class-specific precise feature representations.Extensive experiments on public datasets,including COCO2014,VOC2007,and VOC2012,demonstrate the effectiveness of our proposed method.Our approach achieves consistent performance gains of 0.3%,0.4%,and 0.2%over state-of-the-art methods on the three datasets,respectively.These results validate the reliability and superiority of our approach for multi-label image classification.
文摘Because the physiological characteristics and melanin regulation mechanism of zebrafish are highly similar with those of humans,it is of high reference value to use zebrafish model in the evaluation of cosmetic whitening efficacy.In this study,zebrafish embryos are used as biological models to evaluate the whitening efficacy of six kinds of cosmetics raw materials,such as antioxidant,preservative and essence,and the formula of facial cleanser and facial mask products,and the limitations of the zebrafish melanin production grayscale detection method in practical application are discussed.The results show that the selection of different types of components can also reduce the production of melanin and show whitening effect.It can be seen that the gray scale method of melanin production in zebrafish is suitable for the evaluation of the efficacy of raw materials.In practical application,due to the complexity of the formula,the toxic effects of different types of ingredients may interfere with the melanin generation of zebrafish,affecting the judgment and evaluation of whitening efficacy.For the detection of whitening efficacy of products,a comprehensive evaluation system should be built together with other methods to accurately evaluate the whitening efficacy.
基金supported by the National Key R&D Program of China(No.2022YFB3205101)NSAF(No.U2230116)。
文摘To improve image quality under low illumination conditions,a novel low-light image enhancement method is proposed in this paper based on multi-illumination estimation and multi-scale fusion(MIMS).Firstly,the illumination is processed by contrast-limited adaptive histogram equalization(CLAHE),adaptive complementary gamma function(ACG),and adaptive detail preserving S-curve(ADPS),respectively,to obtain three components.Then,the fusion-relevant features,exposure,and color contrast are selected as the weight maps.Subsequently,these components and weight maps are fused through multi-scale to generate enhanced illumination.Finally,the enhanced images are obtained by multiplying the enhanced illumination and reflectance.Compared with existing approaches,this proposed method achieves an average increase of 0.81%and 2.89%in the structural similarity index measurement(SSIM)and peak signal-to-noise ratio(PSNR),and a decrease of 6.17%and 32.61%in the natural image quality evaluator(NIQE)and gradient magnitude similarity deviation(GMSD),respectively.
基金supported by the National Natural Science Foundation of China(No.42207175)。
文摘The joint roughness coefficient(JRC) is one of the key parameters for evaluating the shear strength of rock joints.Because of the scale effect in the JRC,reliable JRC values are of great importance for most rock engineering projects.During the collection process of JRC samples,the redundancy or insufficiency of representative rock joint surface topography(RJST) information in serial length JRC samples is the essential reason that affects the reliability of the scale effect results.Therefore,this paper proposes an adaptive sampling method,in which we use the entropy consistency measure Q(a) to evaluate the consistency of the joint morphology information contained in adjacent JRC samples.Then the sampling interval is automatically adjusted according to the threshold Q(at) of the entropy consistency measure to ensure that the degree of change of RJST information between JRC samples is the same,and ultimately makes the representative RJST information in the collected JRC samples more balanced.The application results of actual cases show that the proposed method can obtain the scale effect in the JRC efficiently and reliably.
基金funded by the Deanship of Scientific Research,Princess Nourah bint Abdulrahman University,through the Program of Research Project Funding After Publication,grant No(44-PRFA-P-107).
文摘This study reports the analytical solution for a generalized rotational pendulum system with gallows and periodic excited forces.The multiple scales method(MSM)is applied to solve the proposed problem.Several types of rotational pendulum oscillators are studied and talked about in detail.These include the forced damped rotating pendulum oscillator with gallows,the damped standard simple pendulum oscillator,and the damped rotating pendulum oscillator without gallows.The MSM first-order approximations for all the cases mentioned are derived in detail.The obtained results are illustrated with concrete numerical examples.The first-order MSM approximations are compared to the fourth-order Runge-Kutta(RK4)numerical approximations.Additionally,the maximum error is estimated for the first-order approximations obtained through the MSM,compared to the numerical approximations obtained by the RK4 method.Furthermore,we conducted a comparative analysis of the outcomes obtained by the used method(MSM)and He-MSM to ascertain their respective levels of precision.The proposed method can be applied to analyze many strong nonlinear oscillatory equations.
基金National Natural Science Foundation of China (2009ZB52028,05C52013)Ph.D. Programs Foundation of Ministry of Education of China (20070287039)
文摘In the analysis of functionally graded materials （FGMs）, the uncoupled approach is used broadly, which is based on homogenized material property and ignores the effect Of local micro-structural interaction. The higher-order theory for FGMs （HOTFGM） is a coupled approach that explicitly takes the effect of micro-structural gradation and the local interaction of the spatially variable inclusion phase into account. Based on the HOTFGM, this article presents a quadrilateral element-based method for the calculation of multi-scale temperature field （QTF）. In this method, the discrete cells are quadrilateral including rectangular while the surface-averaged quantities are the primary variables which replace the coefficients employed in the temperature function. In contrast with the HOTFGM, this method improves the efficiency, eliminates the restriction of being rectangular cells and expands the solution scale. The presented results illustrate the efficiency of the QTF and its advantages in analyzing FGMs.
基金Project supported by the National Natural Science Foundation of China （Grant No.10472060）
文摘The natural frequencies of an axially moving beam were determined by using the method of multiple scales. The method of second-order multiple scales could be directly applied to the governing equation if the axial motion of the beam is assumed to be small. It can be concluded that the natural frequencies affected by the axial motion are proportional to the square of the velocity of the axially moving beam. The results obtained by the perturbation method were compared with those given with a numerical method and the comparison shows the correctness of the multiple-scale method if the velocity is rather small.
基金supported by Laboratory Directed Research and Development(LDRD)funding from Berkeley Labsupported by Open Fund of the State Key Laboratory of Geomechanics and Geotechnical Engineering,Institute of Rock and Soil Mechanics,Chinese Academy of Sciences(Grant No.Z017004)。
文摘The greatest challenges of rigorously modeling coupled hydro-mechanical(HM)processes in fractured geological media at different scales are associated with computational geometry.These challenges include dynamic shearing and opening of intersecting fractures at discrete fracture scales as a result of coupled processes,and contact alteration along rough fracture surfaces that triggers structural and physical changes of fractures at micro-asperity scale.In this paper,these challenges are tackled by developing a comprehensive modeling approach for coupled processes in fractured geological media based on numerical manifold method(NMM)at multiple scales.Based on their distinct geometric features,fractures are categorized into three different scales:dominant fracture,discrete fracture,and discontinuum asperity scales.Here the scale is relative,that of the fracture relative to that of the research interest or domain.Different geometric representations of fractures at different scales are used,and different governing equations and constitutive relationships are applied.For dominant fractures,a finite thickness zone model is developed to treat a fracture as a porous nonlinear domain.Nonlinear fracture mechanical behavior is accurately modeled with an implicit approach based on strain energy.For discrete fractures,a zero-dimensional model was developed for analyzing fluid flow and mechanics in fractures that are geometrically treated as boundaries of the rock matrix.With the zero-dimensional model,these fractures can be modeled with arbitrary orientations and intersections.They can be fluid conduits or seals,and can be open,bonded or sliding.For the discontinuum asperity scale,the geometry of rough fracture surfaces is explicitly represented and contacts involving dynamic alteration of contacts among asperities are rigorously calculated.Using this approach,fracture alteration caused by deformation,re-arrangement and sliding of rough surfaces can be captured.Our comprehensive model is able to handle the computational challenges with accurate representation of intersections and shearing of fractures at the discrete fracture scale and rigorously treats contacts along rough fracture surfaces at the discontinuum asperity scale.With future development of three-dimensional(3D)geometric representation of discrete fracture networks in porous rock and contacts among multi-body systems,this model is promising as a basis of 3D fully coupled analysis of fractures at multiple scales,for advancing understanding and optimizing energy recovery and storage in fractured geological media.
基金The project was supported by the National Natural Science Foundation of China(Grant No.42204122).
文摘There are abundant igneous gas reservoirs in the South China Sea with significant value of research,and lithology classification,mineral analysis and porosity inversion are important links in reservoir evaluation.However,affected by the diverse lithology,complicated mineral and widespread alteration,conventional logging lithology classification and mineral inversion become considerably difficult.At the same time,owing to the limitation of the wireline log response equation,the quantity and accuracy of minerals can hardly meet the exploration requirements of igneous formations.To overcome those issues,this study takes the South China Sea as an example,and combines multi-scale data such as micro rock slices,petrophysical experiments,wireline log and element cutting log to establish a set of joint inversion methods for minerals and porosity of altered igneous rocks.Specifically,we define the lithology and mineral characteristics through core slices and mineral data,and establish an igneous multi-mineral volumetric model.Then we determine element cutting log correction method based on core element data,and combine wireline log and corrected element cutting log to perform the lithology classification and joint inversion of minerals and porosity.However,it is always difficult to determine the elemental eigenvalues of different minerals in inversion.This paper uses multiple linear regression methods to solve this problem.Finally,an integrated inversion technique for altered igneous formations was developed.The results show that the corrected element cutting log are in good agreement with the core element data,and the mineral and porosity results obtained from the joint inversion based on the wireline log and corrected element cutting log are also in good agreement with the core data from X-ray diffraction.The results demonstrate that the inversion technique is applicable and this study provides a new direction for the mineral inversion research of altered igneous formations.
文摘The solution to heat transfer problems in two-dimensional heterogeneous media is attended based on the scaled boundary finite element method(SBFEM)coupled with equilibrated basis functions(EqBFs).The SBFEM reduces the model order by scaling the boundary solution onto the inner element.To this end,tri-lateral elements are emanated from a scaling center,followed by the development of a semi-analytical solution along the radial direction and a finite element solution along the circumferential/boundary direction.The discretization is thus limited to the boundaries of the model,and the semi-analytical radial solution is found through the solution of an eigenvalue problem,which restricts the methods’applicability to heterogeneous media.In this research,we first extracted the SBFEM formulation considering the heterogeneity of the media.Then,we replaced the semi-analytical radial solution with the EqBFs and removed the eigenvalue solution step from the SBFEM.The varying coefficients of the partial differential equation(PDE)resulting from the heterogeneity of the media are replaced by a finite series in the radial and circumferential directions of the element.A weighted residual approach is applied to the radial equation.The equilibrated radial solution series is used in the new formulation of the SBFEM.
基金supported by the National Key R&D Program of China(Grant No.2022YFB3303500).
文摘The present study proposes a sub-grid scale model for the one-dimensional Burgers turbulence based on the neuralnetwork and deep learning method.The filtered data of the direct numerical simulation is used to establish thetraining data set,the validation data set,and the test data set.The artificial neural network(ANN)methodand Back Propagation method are employed to train parameters in the ANN.The developed ANN is applied toconstruct the sub-grid scale model for the large eddy simulation of the Burgers turbulence in the one-dimensionalspace.The proposed model well predicts the time correlation and the space correlation of the Burgers turbulence.
基金supported by Hunan Provincial Natural Science Foundation General Project,grant number 2023JJ306932023 Central South University Education and Teaching Reform Research Project,grant number 2023jy045-3Hunan Provincial Key Laboratory of Low Carbon Healthy Building
文摘The construction of an ecological security pattern(ESP)is an effective approach to addressing ecological concerns in urban agglomerations.However,previous research has often overlooked the ecological characteristics at different spatial scales,making it difficult to effectively tackle environmental issues.This study focuses on the Changsha–Zhuzhou–Xiangtan urban agglomeration(CZXUA)to construct an ESP from 2 scales:the urban agglomeration and the metropolitan area.First,ecological sources at different scales were identified and optimized using morphological spatial pattern analysis,ecosystem service supply assessment,landscape connectivity evaluation,and ecological demand intensity analysis.Next,spatial principal component analysis was employed to analyze resistance factors across different spatial scales,enabling the construction of a multi-scale landscape resistance surface.Finally,multi-scale ecological corridors were extracted using circuit theory,and the ecological corridors passing through the central urban area were optimized.As a result,176 ecological sources,337 ecological corridors,60 ecological pinch points,and 30 ecological barrier points were identified,forming the overall ESP for the CZXUA.Furthermore,an ecological protection and optimization pattern,described as“One River,Seven Waters,One Heart,and Eight Wedges”,was proposed based on the ESP.This pattern highlights pivotal areas for ecological protection and restoration within the national land space.This study provides a comprehensive example of constructing a multi-scale ecological corridor network and identifying critical areas for ecological protection and restoration.The findings can serve as a reference for special ecological protection planning in urban agglomerations.
基金the Deputyship for Research&Innovation,Ministry of Education in Saudi Arabia for funding this research work through the project number RI-44-0143
文摘In this investigation,some different approaches are implemented for analyzing a generalized forced damped complex Duffing oscillator,including the hybrid homotopy perturbation method(H-HPM),which is sometimes called the Krylov-Bogoliubov-Mitropolsky(KBM)method and the multiple scales method(MSM).All mentioned methods are applied to obtain some accurate and stable approximations to the proposed problem without decoupling the original problem.All obtained approximations are discussed graphically using different numerical values to the relevant parameters.Moreover,all obtained approximate solutions are compared with the 4thorder Runge-Kutta(RK4)numerical approximation.The maximum residual distance error(MRDE)is also estimated,in order to verify the high accuracy of the obtained analytic approximations.
基金supported by the Integrated Rail Transit Dispatch Control and Intermodal Transport Service Technology Project(Grant No.2022YFB4300500).
文摘The Balise Transmission Module(BTM)unit of the on-board train control system is a crucial component.Due to its unique installation position and complex environment,this unit has a higher fault rate within the on-board train control system.To conduct fault prediction for the BTM unit based on actual fault data,this study proposes a prediction method combining reliability statistics and machine learning,and achieves the fusion of prediction results from different dimensions through multi-method interactive validation.Firstly,a method for predicting equipment fault time targeting batch equipment is introduced.This method utilizes reliability statistics to construct a model of the remaining faultless operating time distribution considering uncertainty,thereby predicting the remaining faultless operating probability of the BTM unit.Secondly,considering the complexity of the BTM unit’s fault mechanism,the small sample size of fault cases,and the potential presence of multiple fault features in fault text records,an individual-oriented fault prediction method based on Bayesian-optimized Gradient Boosting Regression Tree(Bayes-GBRT)is proposed.This method achieves better prediction results compared to linear regression algorithms and random forest regression algorithms,with an average absolute error of only 0.224 years for predicting the fault time of this type of equipment.Finally,a multi-method interactive validation approach is proposed,enabling the fusion and validation of multi-dimensional results.The results indicate that the predicted fault time and the actual fault time conform to a log-normal distribution,and the parameter estimation results are basically consistent,verifying the accuracy and effectiveness of the prediction results.The above research findings can provide technical support for the maintenance and modification of BTM units,effectively reducing maintenance costs and ensuring the safe operation of high-speed railway,thus having practical engineering value for preventive maintenance.
文摘Ceramic Matrix Composite （CMC） turbine guide vanes possess multi-scale stress and strain with inhomogeneity at the microscopic scale. Given that the macroscopic distribution cannot reflect the microscopic stress fluctuation, the macroscopic method fails to meet the requirements of stress and strain analysis of CMC turbine guide vanes. Furthermore, the complete thermodynamic properties of 2D woven SiC/SiC-CMC cannot be obtained through experimentation, Accordingly, a method to calculate the thermodynamic properties of CMC and analyze multi-scale stress and strain of the turbine guide vanes should be established. In this study, the multi-scale thermodynamic analysis is investigated. The thermodynamic properties of Chemical Vapor Infiltration （CVI） pro- cessed SiC/SiC-CMC are predicted by a Representative Volume Element （RVE） model with porosity, leading to the result that the relative error between the calculated in-plane tensile modulus and the experimental value is 4.2%. The macroscopic response of a guide vane under given conditions is predicted. The relative error between the predicted strain on the trailing edge and the experimental value is 9.7%. The calculation of the stress distribution of micro-scale RVE shows that the maximum value of microscopic stress, which is located in the interlayer matrix, is more than 1.5 times that of macroscopic stress in the same direction and the microscopic stress distribution of the interlayer matrix is related to the pore distribution of the composite.
基金Supported by the National Natural Science Foundation of China(51105195,51075204)the Aeronautical Science Foundation of China(2011ZB52024)
文摘A straightforward multi-scale boundary element method is proposed for global and local mechanical analysis of heterogeneous material.The method is more accurate and convenient than finite element based multi-scale method.The formulations of this method are derived by combining the homogenization approach and the fundamental equations of boundary element method.The solution gives the convenient formulations to compute global elastic constants and the local stress field.Finally,two numerical examples of porous material are presented to prove the accuracy and the efficiency of the proposed method.The results show that the method does not require the iteration to obtain the solution of the displacement in micro level.
文摘Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.
