Narrow Results

This paper studies the color expression method of fine arts painting based on multi-scale Retinex, so as to improve the color expression ability of fine arts painting works in digital application. The images of art paintings are converted into HSI and LAB color spaces, respectively. The images are binarized using the Otsu threshold method in the HSI color space, and segmented using the $K$-means clustering algorithm in the LAB color space. The image processing results from the two color spaces are mathematically combined (or calculated) to complete the segmentation of the art painting image. The multi-scale Retinex algorithm is used to map the tone of the segmented image. Based on the trichromatic theory, the color correction is carried out on the image after tone mapping, and the final color expression result of art painting is obtained. The experimental results show that the image processed by this method has high definition, rich color and detail information, and strong overall layering, which makes the color expression of art paintings more ideal.

In this paper, we focus on denoising seismic signals effectively to obtain high-quality data, which is crucially important for oil-gas reservoir prediction and seismic interpretation tasks. The rapid progress of deep learning has brought new development opportunities to seismic oil and gas exploration technologies. However, current deep learning-based seismic denoising models have limited learning ability due to insufficient extraction features in strong noise backgrounds. For this reason, we develop a new multi-scale generative adversarial networks (GAN) with transform prediction for seismic signal denoising. First, we develop a deep multi-scale diversion fusion (MSDF) network in a GAN generator considering the advantages of combining GAN and convolutional neural networks. MSDF network primarily contains several MSDF blocks that mine abundant long-short path features in multiple receptive fields to restore seismic signals in a rough to detailed manner. Additionally, current deep learning-based seismic denoising models only exploit features in the spatial domain, not considering high-frequency characteristics, which results in insufficient high-frequency details when reconstructing seismic data. So, we propose to predict high-frequency components during learning by employing the superior non-subsampled contourlet transform (NSCT), which further preserves the better global topological structure and local texture characteristics of MSDF in GAN generator than the spatial domain, promoting the discrimination ability in GAN discriminator. The qualitative and quantitative results on our constructed synthetic dataset and actual seismic data demonstrate that the proposed method surpasses other deep learning-based approaches in realizing higher signal-to-noise ratio, as well as mining more effective high-frequency signals.

The study addresses the global impact of COVID-19 by developing a mathematical model that combines within-host and between-host factors to better understand the disease’s dynamics. It begins by describing SARS-CoV-2 dynamics within individual human hosts using fractional-order differential equations. The model is shown to be Ulam–Hyers stable, ensuring reliable predictions. The research then investigates virus transmission from infected to susceptible individuals using agent-based modeling (ABM). This approach allows us to capture the diversity and heterogeneity among individuals, including variations in internal state of individuals, immune response and responses to interventions, making the model more realistic compared to aggregate models. The agent-based model places individuals on a square lattice, assigns health states (susceptible, infectious, or recovered), and relies on infected individuals’ viral load for transmission. Parameter values are stochastically generated via Latin hypercube sampling. The study further explores the impact of viral mutation and control measures. Simulations demonstrate that vaccination substantially reduces transmission but may not eliminate it entirely. The strategy is more effective when vaccinated individuals are evenly distributed across the population, as opposed to concentrated on one side. The research further reveals that while reducing transmission probability decreases infections by implementing prevention protocols, it does not proportionally correlate with the reduction magnitude. This discrepancy is attributed to the intervention primarily addressing inter-host transmission dynamics without directly influencing intra-host viral dynamics.

To address the complex and variable attack patterns prevalent in the current cybersecurity landscape, as well as the limitations of traditional detection methods concerning sensing range, accuracy and efficiency, an innovative intelligent detection model for cybersecurity was proposed. First, we design a distributed multi-target finite sensing mechanism with complementary fields of view, significantly extending the sensing range by optimizing sensor layout and collaboration strategies. Second, this study constructs a multi-scale attention network model (MSANet), which enhances the feature extraction and expression capabilities of the model without imposing additional computational burdens, thereby enabling more accurate recognition of cyber-attack patterns across different scales. Finally, leveraging the extensive data provided by the distributed perception system and the robust learning capabilities of MSANet, we develop a label-free intelligent detection model for network security. This model effectively addresses the detection challenges arising from feature distribution discrepancies between the target and source network domains, achieving efficient and accurate detection in environments with no labeled or limited labeled data. This advancement provides substantial technical support for network security protection. Experimental results demonstrate that our approach achieves accuracy rates of 98.2%, 96.7% and 99.5%, as well as F1-scores of 97.9%, 95.8% and 97.9%, respectively, in detecting botnet traffic, background traffic and normal traffic within the CTU network traffic dataset. In summary, this study not only enriches the theoretical framework of network security detection but also offers practical solutions for constructing an efficient and intelligent network security protection system, possessing significant theoretical value and promising application prospects.

In this study, multi-scale creep analysis of plain-woven GFRP laminates is performed using the time-dependent homogenization theory developed by the present authors. First, point-symmetry of internal structures of plain-woven laminates is utilized for a boundary condition of unit cell problems, reducing the domain of analysis to 1/4 and 1/8 for in-phase and out-of-phase laminate configurations, respectively. The time-dependent homogenization theory is then reconstructed for these domains of analysis. Using the present method, in-plane creep behavior of plain-woven glass fiber/epoxy laminates subjected to a constant stress is analyzed. The results are summarized as follows: (1) The in-plane creep behavior of the plain-woven GFRP laminates exhibits marked anisotropy. (2) The laminate configurations considerably affect the creep behavior of the laminates.

In this paper, a novel approach derived from image gradient domain called multi-scale gradient faces (MGF) is proposed to abstract multi-scale illumination-insensitive measure for face recognition. MGF applies multi-scale analysis on image gradient information, which can discover underlying inherent structure in images and keep the details at most while removing varying lighting. The proposed approach provides state-of-the-art performance on Extended YaleB and PIE: Recognition rates of 99.11% achieved on PIE database and 99.38% achieved on YaleB which outperforms most existing approaches. Furthermore, the experimental results on noised Yale-B validate that MGF is more robust to image noise.

In the original compression tracking algorithm, the size of the tracking box is fixed. There should be better tracking results for scale-invariant objects, but worse tracking results for scale-variant objects. To overcome this defect, a scale-adaptive compressive tracking (CT) algorithm is proposed. First of all, the imbalance of the gray and texture features in the original CT algorithm is balanced by the multi-feature method, which makes the algorithm more robust. Then, searching different candidate regions by using the method of multi-scale search along with feature normalization makes the features extracted from different scales comparable. Finally, the candidate region with the maximum discriminate degree is selected as the object region. Thus, the tracking-box size is adaptive. The experimental results show that when the object scale changes, the improving CT algorithm has higher accuracy and robustness than the original CT algorithm.

Simultaneous extraction of spectral and spatial features and their fusion is currently a popular solution in hyperspectral image (HSI) classification. It has achieved satisfactory results in some research. Because the scales of objects are often different in HSI, it is necessary to extract multi-scale features. However, this aspect was not taken into account in many spectral-spatial feature fusion methods. This causes the model to be unable to get sufficient features on scales with a large difference range. The model (MCMN: Multi-Content Merging Network) proposed in this paper designs a multi-branch fusion structure to extract multi-scale spatial features by using multiple dilated convolution kernels. Considering the interference of the surrounding heterogeneous objects, the useful information from different directions is also fused together to realize the merging of multiple regional features. MCMN introduces a convolution block attention mechanism, which fully extracts attention features in both spatial and spectral directions, so that the network can focus on more useful parts, which can effectively improve the performance of the model. In addition, since the number of objects in each class is often discrepant, it will have some impact on the training process. We apply the focal loss function to eliminate the negative factor. The experimental results of MCMN on three data sets have a breakthrough compared with the other comparison models, which highlights the role of MCMN structure.

To improve the accuracy of cross-domain object detection, the existing unsupervised domain adaptation (UDA) object detection methods mostly use Feature Pyramid Network (FPN), multiple Region Proposal Network (RPN), and multiple domain classifier, but these methods lead to complex network structures, slow model convergence, and low detection efficiency. To solve the above problems, this paper proposes an Efficient and Accurate Cross-domain Object Detection Method Using One-level Feature and Domain Adaptation (OFDA). This method realizes one-level feature object detection through feature fusion and divide-and-conquer technology; realizes overfitting feature suppression and unsupervised domain adaptation through domain-specific suppression and domain feature alignment technology; realizes background feature suppression through the Objectness branch, which replaces the time-consuming Region Proposal Network (RPN) structure and improves the efficiency of unsupervised adaptive detection. The paper verifies the feasibility and superiority of the proposed method by the comparative experiments and ablation experiments of multiple datasets. The proposed OFDA method not only improves the efficiency of object detection, but also ensures the accuracy of cross-domain detection.

Challenges such as complex backgrounds, drastic variations in target scales, and dense distributions exist in natural scenes. Some algorithms optimize multi-scale object detection performance by combining low-level and high-level information through feature fusion strategies. However, these methods overlook the inherent spatial properties of objects and the relationships between foreground and background. To fundamentally enhance the multi-scale detection capability, we propose a depth-constrained multi-scale object detection network that simultaneously learns object detection and depth estimation through a unified framework. In this network, depth features are merged into the detection branch as auxiliary information and constrained and guided to obtain better spatial representations, which enhances discrimination between multi-scale objects. We also introduce a novel cross-modal fusion (CmF) strategy that utilizes depth awareness and low-level detail clues to supplement edge information and adjust attention weight preferences. We find complementary information from RGB and high-quality depth features to achieve better multi-modal information fusion. Experimental results demonstrate that our method outperforms state-of-the-art methods on the KINS dataset, with an improvement of 3.0% in AP score over the baseline network. Furthermore, we validate the effectiveness of our proposed method on the KITTI dataset.

At present, footprint image retrieval based on deep learning mainly focuses on complete footprint, but many of the footprints obtained in the field of public safety and criminal investigation are incomplete forms, Therefore, the feature analysis of incomplete footprint has important practical significance. Based on multiple scale features fusion, we proposed a method to solve incomplete footprints retrieval. On the basis of extracting the global feature of footprint, this method simultaneously extracts multiple local features from different stages of the backbone network to supplement the footprint feature information. The multi-scale feature orthogonal fusion module is used to reduce redundant features, improve the expression ability of footprint features, and solve the problem of incomplete footprint retrieval to a certain extent. The experiment shows that our method has certain effectiveness in retrieving problems on incomplete footprint, with a Top 1 accuracy of 87.08%, expanding the scope of footprint research.

In modern convolutional neural network (CNN)-based object detector, the extracted features are not suitable for multi-scale detection and all the bounding boxes are simply ranked according to their classification scores in nonmaximum suppression (NMS). To address the above problems, we propose a novel one-stage detector named receptive field fusion RetinaNet. First, receptive field fusion module is proposed to extract richer multi-scale features by fusing feature maps of various receptive fields. Second, joint confidence guided NMS is proposed to optimize the post-processing process of object detection, which introduce location confidence in NMS and take joint confidence as the NMS rank basis. According to our experimental results, significant improvement in terms of mean of average precision (mAP) can be achieved on average compared with the state-of-the-art algorithm.

Leptospirosis is a bacteria infection prevalent in many tropical regions, caused by the genus Leptospira. Humans contract the disease by coming in contact with contaminated environments. This study proposes a deterministic mathematical model that links the within-host and between-host dynamics of leptospirosis and investigates its properties. The model’s parameters were estimated by fitting it to real-life data using the “lsqcurvefit” package in MATLAB. The study employs global sensitivity analysis using Latin hypercube sampling with a partial rank correlation coefficient index and uses Pontryagin’s maximum principle to identify cost-effective solutions for time-dependent intervention strategies to suppress the bacteria transmission within a specific period. The results of the study showed that bacteria replication within human and rodent hosts is the major driver of the overall dynamics of leptospirosis and therefore controlling the prevalence of the disease by focusing on the epidemiology components is necessary but will not be effective if the intra-host dynamics of the bacteria within rodent hosts and human hosts are not attentively considered.

Convective heat transfer process between fractal tube bank and coolant water was simulated by using the finite element method. Both rectangular and circular tube banks were investigated in detail. For the relationship between Nusselt number and Reynolds number, the fractal circular tube bank shows the same trend as the rectangular tube one. Based on simulation results, the empirical heat transfer correlations for different stage fractals were regressed. The exponents of correlations increase with the fractal stage. Almost linear correlations between Nu and Re for fourth stage fractal were obtained, indicating a good heat transfer enhancement with the increase in fractal stages. Compared to the uniform tube bank, the fractal tube banks show better overall performance in enhanced heat transfer effect. This research would be helpful to understand the heat exchange mechanism of the multi-scale structure.

The resolution of structural finite element model (FEM) determines the computation cost and accuracy in dynamic analysis. This study proposes a novel wavelet finite element model (WFEM), which facilitates adaptive mesh refinement, for the dynamic analysis and damage detection of beam structures subjected to a moving load (ML). The multi-scale equations of motion for the beam under the ML are derived using the second-generation cubic Hermite multi-wavelets as the shape functions. Then an adaptive-scale analysis strategy is established, in which the scales of the wavelet beam elements are dynamically changed according to the ML position. The performance of the multi-scale WFEM is examined in both dynamic analysis and damage detection problems. It is demonstrated that the multi-scale WFEM with a similar number of degrees of freedom can achieve much higher accuracy than the traditional FEM. In particular, the multi-scale WFEM enables the detection of sub-element damage with a progressive model updating process. The advantage in computation efficiency and accuracy makes the proposed method a promising tool for multi-scale dynamic analysis or damage detection of structures.

The musculoskeletal system, containing bones, cartilage, skeletal muscles, tendons, ligaments, and some other tissues, is a perfect system that undergoes the external and internal load properly and controls the body’s motion efficiently. In this system, skeletal muscle is obviously indispensable. People have been studying the mystery of skeletal muscle mechanics for the last 80 years. Many modeling methods have been used to study skeletal muscles. Among these methods, multi-scale modeling methods are increasingly frequently used in studying musculoskeletal systems, especially those of skeletal muscles. In this review, we summarize the multi-scale modeling methods in studying works of skeletal muscle modeling reported so far. Then, several multi-scale methods of other tissues which possibly could be used in research on skeletal muscle modeling are discussed. Finally, the future research direction and the main challenges of multi-scale skeletal muscle modeling are briefly presented.

As the aging population continues to grow, the increasing number of disabled elderly individuals poses ongoing challenges to society in terms of healthcare and medical burdens. Grading disability is of significant importance for improving medical services and allocating social resources. Traditional grading methods using a single assessment scale tend to be subjective and cannot comprehensively and accurately reflect the real situation. This paper presents the NES (Needs assessment of Elderly care in Shanghai) scale, based on 33 quantified scoring items from basic surveys, to enhance the objectivity and accuracy of disability grading. Correlation analysis of observed items using the C4.5 decision tree algorithm is conducted, followed by an analysis and evaluation of the scale’s performance. The data used in this study are derived from 266 questionnaires collected from eight regions in Shanghai and categorized into three classes: physiological (JKOM/BADL), psychological (BDI/SDS/SAS), and cognitive (CDR). The analysis results show that the NES scale exhibits better robustness in physiological assessment compared to the BADL and JKOM scales. In psychological assessment, the NES scale performs similarly to the SDS and SAS scales. In cognitive assessment, the NES scale’s grading level is slightly higher than that of the CDR scale, demonstrating good comprehensive disability assessment capabilities.

The acquisition and transmission of magnetic resonance (MR) images are frequently affected by random noise pollution, which hampers the diagnosis of diseases by doctors or automated systems. Hence, the search for advanced denoising methods is of great research interest, particularly in magnetic resonance imaging (MRI) denoising models, which are based on deep learning networks and achieve satisfactory results. However, the mining of noisy contextual information and effective information-guided transfer are often neglected in the denoising process, which leads to poor extraction of information at different scales, and poor retention of details. This greatly hinders the further development of MR image denoising methods. Here, we propose a denoising method, MSDRA-Net, for the mining and exploitation of different hierarchical noise features and construct a multi-scale dilated residual (MSDR) structure to transfer and retain noise information at different levels across the layer. Next, a contextual guidance attention (CGA) module guides and transfers contextual information, utilizing the features learnt from different layers of the model as weights. A reconstruction refinement block (RRB) is utilized to construct clean images from the obtained noise bias and the given noisy images. Experiments on simulated and clinical MRI data validated the effectiveness of our method, which demonstrated a superior performance compared to several state-of-the-art methods.

The main mechanisms that control the organization of multicellular tissues are still largely open. A commonly used tool to study basic control mechanisms are in vitro experiments in which the growth conditions can be widely varied. However, even in vitro experiments are not free from unknown or uncontrolled influences. One reason why mathematical models become more and more a popular complementary tool to experiments is that they permit the study of hypotheses free from unknown or uncontrolled influences that occur in experiments. Many model types have been considered so far to model multicellular organization ranging from detailed individual-cell based models with explicit representations of the cell shape to cellular automata models with no representation of cell shape, and continuum models, which consider a local density averaged over many individual cells. However, how the different model description may be linked, and, how a description on a coarser level may be constructed based on the knowledge of the finer, microscopic level, is still largely unknown. Here, we consider the example of monolayer growth in vitro to illustrate how, in a multi-step process starting from a single-cell based off-lattice-model that subsumes the information on the sub-cellular scale by characteristic cell-biophysical and cell-kinetic properties, a cellular automaton may be constructed whose rules have been chosen based on the findings in the off-lattice model. Finally, we use the cellular automaton model as a starting point to construct a multivariate master equation from a compartment approach from which a continuum model can be derived by a systematic coarse-graining procedure. We find that the resulting continuum equation largely captures the growth behavior of the CA model. The development of our models is guided by experimental observations on growing monolayers.

Domain switch, phase transition and electrochemical mechanisms introduced traditional smart materials to various technologies such as piezoelectric ceramics, shape memory alloys and elastor/polymer used in actuators, sensors, etc. From the studies on carbon nanotues, graphite and other nanoscaled materials, we show that due to the quantum effect induced response at atomic or molecular level, all matters in this scale would exhibit intelligent behaviors, and buildings based on this are expected to construct nanointelligent systems which can be believed to improve the current technologies.