Systematic Inference of Multi-scale Chromatin Sub-compartments Using Calder2.

The compartmentalization of chromatin reflects its underlying biological activities. Inferring chromatin sub-compartments using Hi-C data is challenged by data resolution constraints. Consequently, comprehensive characterizations of sub-compartments have been limited to a select number of Hi-C experiments, with systematic comparisons across a wide range of tissues and conditions still lacking. Our original Calder algorithm marked a significant advancement in this field, enabling the identification of multi-scale sub-compartments at various data resolutions and facilitating the inference and comparison of chromatin architecture in over 100 datasets. Building on this foundation, we introduce Calder2, an updated version of Calder that brings notable improvements. These include expanded support for a wider array of genomes and organisms, an optimized bin size selection approach for more accurate chromatin compartment detection, and extended support for input and output formats. Calder2 thus stands as a refined analysis tool, significantly advancing genome-wide studies of 3D chromatin architecture and its functional implications.

Multiple diffusion models-enhanced extremely limited-view reconstruction strategy for photoacoustic tomography boosted by multi-scale priors.

Photoacoustic tomography (PAT) is an innovative biomedical imaging technology, which has the capacity to obtain high-resolution images of biological tissue. In the extremely limited-view cases, traditional reconstruction methods for photoacoustic tomography frequently result in severe artifacts and distortion. Therefore, multiple diffusion models-enhanced reconstruction strategy for PAT is proposed in this study. Boosted by the multi-scale priors of the sinograms obtained in the full view and the limited-view case of 240°, the alternating iteration method is adopted to generate data for missing views in the sinogram domain. The strategy refines the image information from global to local, which improves the stability of the reconstruction process and promotes high-quality PAT reconstruction. The blood vessel simulation dataset and the in vivo experimental dataset were utilized to assess the performance of the proposed method. When applied to the in vivo experimental dataset in the limited-view case of 60°, the proposed method demonstrates a significant enhancement in peak signal-to-noise ratio and structural similarity by 23.08 % and 7.14 %, respectively, concurrently reducing mean squared error by 108.91 % compared to the traditional method. The results indicate that the proposed approach achieves superior reconstruction quality in extremely limited-view cases, when compared to other methods. This innovative approach offers a promising pathway for extremely limited-view PAT reconstruction, with potential implications for expanding its utility in clinical diagnostics.

MobileNet-V2: An Enhanced Skin Disease Classification by Attention and Multi-Scale Features.

The increasing prevalence of skin diseases necessitates accurate and efficient diagnostic tools. This research introduces a novel skin disease classification model leveraging advanced deep learning techniques. The proposed architecture combines the MobileNet-V2 backbone, Squeeze-and-Excitation (SE) blocks, Atrous Spatial Pyramid Pooling (ASPP), and a Channel Attention Mechanism. The model was trained on four diverse datasets such as PH2 dataset, Skin Cancer MNIST: HAM10000 dataset, DermNet. dataset, and Skin Cancer ISIC dataset. Data preprocessing techniques, including image resizing, and normalization, played a crucial role in optimizing model performance. In this paper, the MobileNet-V2 backbone is implemented to extract hierarchical features from the preprocessed dermoscopic images. The multi-scale contextual information is fused by the ASPP model for generating a feature map. The attention mechanisms contributed significantly, enhancing the extraction ability of inter-channel relationships and multi-scale contextual information for enhancing the discriminative power of the features. Finally, the output feature map is converted into probability distribution through the softmax function. The proposed model outperformed several baseline models, including traditional machine learning approaches, emphasizing its superiority in skin disease classification with 98.6% overall accuracy. Its competitive performance with state-of-the-art methods positions it as a valuable tool for assisting dermatologists in early classification. The study also identified limitations and suggested avenues for future research, emphasizing the model's potential for practical implementation in the field of dermatology.

Parameter quantification for oxygen transport in the human brain.

BACKGROUND AND OBJECTIVE: Oxygen is carried to the brain by blood flow through generations of vessels across a wide range of length scales. This multi-scale nature of blood flow and oxygen transport poses challenges on investigating the mechanisms underlying both healthy and pathological states through imaging techniques alone. Recently, multi-scale models describing whole brain perfusion and oxygen transport have been developed. Such models rely on effective parameters that represent the microscopic properties. While parameters of the perfusion models have been characterised, those for oxygen transport are still lacking. In this study, we set to quantify the parameters associated with oxygen transport and their uncertainties. METHODS: Effective parameter values of a continuum-based porous multi-scale, multi-compartment oxygen transport model are systematically estimated. In particular, geometric parameters that capture the microvascular topologies are obtained through statistically accurate capillary networks. Maximum consumption rates of oxygen are optimised to uniquely define the oxygen distribution over depth. Simulations are then carried out within a one-dimensional tissue column and a three-dimensional patient-specific brain mesh using the finite element method. RESULTS: Effective values of the geometric parameters, vessel volume fraction and surface area to volume ratio, are found to be 1.42% and 627 [mm2/mm3], respectively. These values compare well with those acquired from human and monkey vascular samples. Simulation results of the one-dimensional tissue column show qualitative agreement with experimental measurements of tissue oxygen partial pressure in rats. Differences between the oxygenation level in the tissue column and the brain mesh are observed, which highlights the importance of anatomical accuracy. Finally, one-at-a-time sensitivity analysis reveals that the oxygen model is not sensitive to most of its parameters; however, perturbations in oxygen solubilities and plasma to whole blood oxygen concentration ratio have a considerable impact on the tissue oxygenation. CONCLUSIONS: The findings of this study demonstrate the validity of using a porous continuum approach to model organ-scale oxygen transport and draw attention to the significance of anatomy and parameters associated with inter-compartment diffusion.

Multi-scale transcriptomics reveals that specific tumor cells promote lung adenocarcinoma metastasis through crosstalk with the microenvironment.

Most advanced lung adenocarcinoma (LUAD) patient deaths are attributed to metastasis. However, the complete understanding of the metastatic mechanism in LUAD remains elusive. Single-cell RNA-seq (scRNA-seq), spatial RNA-seq (stRNA-seq) and bulk RNA-seq of primary LUAD were integrated to investigate metastatic driver genes, cell-cell interactions, and spatial colocalization of cells and ligand-receptor pairs. A lung adenocarcinoma metastasis risk scoring model (LMRS) was established to estimate the risk of metastasis in LUAD. Forty-two metastasis driver genes were identified and tumor epithelial cells were classified into two subtypes. Epithelial cell subclass characterized by susceptibility to metastasis are referred to as Epithelial_LM, and the remaining as Epithelial_LL. Epithelial_LM subtype has intimate ligand-receptor interactions with inflammatory endothelial cells (iendo), inflammatory cancer-associated fibroblasts (iCAF), and NKT cells. Epithelial_LM cells have a spatial colocalization relationship with these three types of cells. The LMRS was established and its efficacy was verified in bulk RNA-seq. We identified a subclass of epithelial cells prone to metastasis and demonstrated the contribution of inflammatory stromal cells and NKT cells in facilitating tumor metastasis.

A general multi-scale description of metastable adaptive motion across fitness valleys.

We consider a stochastic individual-based model of adaptive dynamics on a finite trait graph G = ( V , E ) . The evolution is driven by a linear birth rate, a density dependent logistic death rate and the possibility of mutations along the directed edges in E. We study the limit of small mutation rates for a simultaneously diverging population size. Closing the gap between Bovier et al. (Ann Appl Probab 29(6):3541-358, 2019) and Coquille et al. (Electron J Probab 26:1-37, 2021) we give a precise description of transitions between evolutionary stable conditions (ESC), where multiple mutations are needed to cross a valley in the fitness landscape. The system shows a metastable behaviour on several divergent time scales, corresponding to the widths of these fitness valleys. We develop the framework of a meta graph that is constituted of ESCs and possible metastable transitions between them. This allows for a concise description of the multi-scale jump chain arising from concatenating several jumps. Finally, for each of the various time scales, we prove the convergence of the population process to a Markov jump process visiting only ESCs of sufficiently high stability.

Lightweight medical image segmentation network with multi-scale feature-guided fusion.

In the field of computer-aided medical diagnosis, it is crucial to adapt medical image segmentation to limited computing resources. There is tremendous value in developing accurate, real-time vision processing models that require minimal computational resources. When building lightweight models, there is always a trade-off between computational cost and segmentation performance. Performance often suffers when applying models to meet resource-constrained scenarios characterized by computation, memory, or storage constraints. This remains an ongoing challenge. This paper proposes a lightweight network for medical image segmentation. It introduces a lightweight transformer, proposes a simplified core feature extraction network to capture more semantic information, and builds a multi-scale feature interaction guidance framework. The fusion module embedded in this framework is designed to address spatial and channel complexities. Through the multi-scale feature interaction guidance framework and fusion module, the proposed network achieves robust semantic information extraction from low-resolution feature maps and rich spatial information retrieval from high-resolution feature maps while ensuring segmentation performance. This significantly reduces the parameter requirements for maintaining deep features within the network, resulting in faster inference and reduced floating-point operations (FLOPs) and parameter counts. Experimental results on ISIC2017 and ISIC2018 datasets confirm the effectiveness of the proposed network in medical image segmentation tasks. For instance, on the ISIC2017 dataset, the proposed network achieved a segmentation accuracy of 82.33 % mIoU, and a speed of 71.26 FPS on 256 × 256 images using a GeForce GTX 3090 GPU. Furthermore, the proposed network is tremendously lightweight, containing only 0.524M parameters. The corresponding source codes are available at https://github.com/CurbUni/LMIS-lightweight-network.

A novel multi-scale network intrusion detection model with transformer.

Network is an essential tool today, and the Intrusion Detection System (IDS) can ensure the safe operation. However, with the explosive growth of data, current methods are increasingly struggling as they often detect based on a single scale, leading to the oversight of potential features in the extensive traffic data, which may result in degraded performance. In this work, we propose a novel detection model utilizing multi-scale transformer namely IDS-MTran. In essence, the collaboration of multi-scale traffic features broads the pattern coverage of intrusion detection. Firstly, we employ convolution operators with various kernels to generate multi-scale features. Secondly, to enhance the representation of features and the interaction between branches, we propose Patching with Pooling (PwP) to serve as a bridge. Next, we design multi-scale transformer-based backbone to model the features at diverse scales, extracting potential intrusion trails. Finally, to fully capitalize these multi-scale branches, we propose the Cross Feature Enrichment (CFE) to integrate and enrich features, and then output the results. Sufficient experiments show that compared with other models, the proposed method can distinguish different attack types more effectively. Specifically, the accuracy on three common datasets NSL-KDD, CIC-DDoS 2019 and UNSW-NB15 has all exceeded 99%, which is more accurate and stable.

Entrapment of multi-scale structure of alginate beads stabilized with cellulose nanofibrils for potential intestinal delivery of lactic acid bacteria.

Soybean cellulose nanofibrils (SCNFs) were formed by autoclave-enzymatic hydrolysis combined with ball milling. SCNFs were blended with sodium alginate (SA) to encapsulate lactic acid bacteria (LAB) through inotropic gelation. The effect of SCNFs on the multiscale structure of SA beads, leading to changes in the survival and release of LAB during simulated digestion, was investigated. Microscopy and rheological testing indicated that SCNF10-30 was well-dispersed in the SA paste in the form of interlaced nanofibrils, and could reduce the deformation of the paste under stress by 47.31 %. Multiscale structural analysis indicated SCNF10-30 not only increased the immobilized water of SA beads by 15.59 % by coordinating calcium, but also regulated the in situ-assembly of SA beads, including an increase in the scale of dimers from 6.73 nm to 8.32 nm and improved arrangement, thus forming a dense gel network. LAB viability of SA-SCNF10-30 in simulated digestion was increased by 1.3 log CFU/g compared to SA beads. Cellulose nanofibrils improved gastrointestinal survival and controlled release of LAB better than fiber rods. This study provides a strategy to regulate the multiscale structure of SA beads through nanofibrils to enable stabilization and sustainable release of LAB in gastrointestinal fluids.

A cluster-based incremental potential approach for reduced order homogenization of bones.

We develop a cluster-based model order reduction (called C-pRBMOR) approach for efficient homogenization of bones, compatible with a large variety of generalized standard material (GSM) models. To this end, the pRBMOR approach based on a mixed incremental potential formulation is extended to a clustered version for a significantly improved computational efficiency. The microscopic modeling of bones falls into a mixed incremental class of the GSM framework, originating from two potentials. An offline phase of the C-pRBMOR approach includes both a clustering analysis spatially decomposing the micro-domain within an RVE and a space-time decomposition of the microscopic plastic strain fields. A comparative study on two different clustering approaches and two algorithms for mode identification is additionally conducted. For an online analysis, a cluster-enhanced version of evolution equations for the reduced variables is derived from an effective incremental variational formulation, rendering a very small set of nonlinear equations to be numerically solved. Several numerical examples show the effectiveness of the C-pRBMOR approach. A striking acceleration rate beyond 104 against conventional FE computations and that beyond 103 against the original pRBMOR approach are observed.

Optimizing secure multimedia communication in embedded systems a parallel convolutional neural network approach to RIS and D2D resource allocation.

With the rapid development of Internet of Things (IoT) services, technologies that leverage multimedia computer communication for information sharing in embedded systems have become a research focus. To address the challenges of low spectral efficiency and poor network flexibility in multimedia computer communications, this paper proposes a resource allocation scheme based on parallel Convolutional Neural Network (CNN). The scheme optimizes the base station beamforming vector and the Reconfigurable Intelligent Surface (RIS) phase shifts to maximize the secure transmission rate for cellular users (CUs), while ensuring normal and secure communication for device-to-device (D2D) users. First, to mitigate interference caused by D2D users reusing CU spectrum resources, the RIS phase shifts and beamforming vectors are optimized to suppress interference and enhance system secrecy rates. Second, to maximize the CU secrecy rate, the paper proposes a parallel CNN-based resource allocation model that considers base station transmission power, RIS reflection coefficients, and D2D communication rate constraints, incorporating multi-scale residual modules in the convolutional layers of the model. Simulation results demonstrate that the proposed CNN-based resource allocation scheme significantly improves the secrecy rate of embedded system communications, ensuring secure multimedia computing, and outperforms traditional methods.

Multi-scale input layers and dense decoder aggregation network for COVID-19 lesion segmentation from CT scans.

Accurate segmentation of COVID-19 lesions from medical images is essential for achieving precise diagnosis and developing effective treatment strategies. Unfortunately, this task presents significant challenges, owing to the complex and diverse characteristics of opaque areas, subtle differences between infected and healthy tissue, and the presence of noise in CT images. To address these difficulties, this paper designs a new deep-learning architecture (named MD-Net) based on multi-scale input layers and dense decoder aggregation network for COVID-19 lesion segmentation. In our framework, the U-shaped structure serves as the cornerstone to facilitate complex hierarchical representations essential for accurate segmentation. Then, by introducing the multi-scale input layers (MIL), the network can effectively analyze both fine-grained details and contextual information in the original image. Furthermore, we introduce an SE-Conv module in the encoder network, which can enhance the ability to identify relevant information while simultaneously suppressing the transmission of extraneous or non-lesion information. Additionally, we design a dense decoder aggregation (DDA) module to integrate feature distributions and important COVID-19 lesion information from adjacent encoder layers. Finally, we conducted a comprehensive quantitative analysis and comparison between two publicly available datasets, namely Vid-QU-EX and QaTa-COV19-v2, to assess the robustness and versatility of MD-Net in segmenting COVID-19 lesions. The experimental results show that the proposed MD-Net has superior performance compared to its competitors, and it exhibits higher scores on the Dice value, Matthews correlation coefficient (Mcc), and Jaccard index. In addition, we also conducted ablation studies on the Vid-QU-EX dataset to evaluate the contributions of each key component within the proposed architecture.

Neuromodulation effect of temporal interference stimulation based on network computational model.

Deep brain stimulation (DBS) has long been the conventional method for targeting deep brain structures, but noninvasive alternatives like transcranial Temporal Interference Stimulation (tTIS) are gaining traction. Research has shown that alternating current influences brain oscillations through neural modulation. Understanding how neurons respond to the stimulus envelope, particularly considering tTIS's high-frequency carrier, is vital for elucidating its mechanism of neuronal engagement. This study aims to explore the focal effects of tTIS across varying amplitudes and modulation depths in different brain regions. An excitatory-inhibitory network using the Izhikevich neuron model was employed to investigate responses to tTIS and compare them with transcranial Alternating Current Stimulation (tACS). We utilized a multi-scale model that integrates brain tissue modeling and network computational modeling to gain insights into the neuromodulatory effects of tTIS on the human brain. By analyzing the parametric space, we delved into phase, amplitude, and frequency entrainment to elucidate how tTIS modulates endogenous alpha oscillations. Our findings highlight a significant difference in current intensity requirements between tTIS and tACS, with tTIS requiring notably higher intensity. We observed distinct network entrainment patterns, primarily due to tTIS's high-frequency component, whereas tACS exhibited harmonic entrainment that tTIS lacked. Spatial resolution analysis of tTIS, conducted via computational modeling and brain field distribution at a 13 Hz stimulation frequency, revealed modulation in deep brain areas, with minimal effects on the surface. Notably, we observed increased power within intrinsic and stimulation bands beneath the electrodes, attributed to the high stimulus signal amplitude. Additionally, Phase Locking Value (PLV) showed slight increments in non-deep areas. Our analysis indicates focal stimulation using tTIS, prompting further investigation into the necessity of high amplitudes to significantly affect deep brain regions, which warrants validation through clinical experiments.

Research on improved YOLOv8s model for detecting mycobacterium tuberculosis.

Accurate identification of Mycobacterium tuberculosis (M. tuberculosis) is a critical step in the diagnosis of tuberculosis. Existing object detection methods struggle with the challenges posed by the varied morphology and size of M. tuberculosis in sputum smear images, which makes precise targeting difficult. To solve these problems, an improved YOLOv8s model is proposed. Specifically, an additional detection head is added to focus on small target information. Second, a multi-scale feature fusion module is introduced to adapt the model to different sizes of M. tuberculosis. In addition, a convolutional layer is added to the Coordinate Attention (CA) module to extract more advanced semantic features. Finally, a self-attention mechanism is added after the CA module to enhance the model's ability to accurately understand and localize the varied morphology of M. tuberculosis. Our model performed well with an average precision of 85.7 % when tested on a publicly available dataset. This clearly demonstrates the effectiveness of our proposed model in M. tuberculosis detection.

Construction of urban-rural ecological networks with multi-scale nesting and composite functions based on the "red-green-blue" spatial perspective: A case study of Dali City, China.

Urbanization has facilitated economic development while simultaneously resulting in various ecological issues. Constructing a multi-scale nested and composite functional urban-rural ecological network is crucial for improving ecological security. This study utilizes Dali City as a case study and employs methods including MSPA, circuit theory, and landscape connectivity index to develop the urban-rural habitat network, water green network, and recreation network, focusing on the " red-green-blue " spatial framework. An analysis of the spatial characteristics of source areas, corridors, ecological strategic points, and other spatial elements is conducted to establish a multi-level, multi-objective, and multifunctional composite urban-rural ecological network. The results show that: (1) 13 ecological source areas were identified in both the municipal and main urban areas, along with 22 ecological corridors in the municipal and 20 main urban areas. The distribution of ecological corridors was uniform across the study area. (2) The optimal width for the municipal biological corridor is 150 m, the main urban area should have a width of 90 m. The optimal width for rainwater corridors in municipal and main urban areas is 60 m. (3) The multi-scale nested ecological network identified 4 common ecological sources, 11 ecological corridors, 3 rainwater corridors, 6 wetland nodes, and 7 amusement nodes. Overall, the number of ecological nodes is limited, indicating a need for enhanced node construction. The research findings offer insights for developing ecological networks that integrate urban and rural functions, serving as a reference for ecological protection and restoration in pertinent regions.

Aggregation-Disruption-Induced Multi-Scale Mediating Strategy for Anticoagulation in Blood-Contacting Devices.

Minimally invasive blood-contacting interventional devices are increasingly used to treat cardiovascular diseases. However, the risk of device-related thrombosis remains a significant concern, particularly the formation of cycling thrombi, which pose life-threatening risks. To better understand the interactions between these devices and blood, the initial stages of coagulation contact activation on extrinsic surfaces are investigated. Direct force measurements reveals that activated contact factors stimulate the intrinsic coagulation pathway and promote surface crosslinking of fibrin. Furthermore, fibrin aggregation is disrupted by surface-grafted inhibitors, as confirmed by ex vivo coagulation tests. An engineered serum protein with zwitterion grafts to resist the deposition of biological species such as fibrin, platelets, and red blood cells is also developed. Simultaneously, a protease inhibitor-based coacervate is incorporated into the coating to inhibit the intrinsic pathway effectively. The loaded coacervate can be released and reloaded through modulation of catechol-amine interactions, facilitating material regeneration. The strategy offers a novel multi-scale mediation strategy that simultaneously inhibits nanoscale coagulation factors and resists microscale thrombus aggregation, providing a long-term solution for anticoagulation in blood-contacting devices.

A Deep Learning Model for Cervical Optical Coherence Tomography Image Classification.

Objectives: Optical coherence tomography (OCT) has recently been used in gynecology to detect cervical lesions in vivo and proven more effective than colposcopy in clinical trials. However, most gynecologists are unfamiliar with this new imaging technique, requiring intelligent computer-aided diagnosis approaches to help them interpret cervical OCT images efficiently. This study aims to (1) develop a clinically-usable deep learning (DL)-based classification model of 3D OCT volumes from cervical tissue and (2) validate the DL model's effectiveness in detecting high-risk cervical lesions, including high-grade squamous intraepithelial lesions and cervical cancer. Method: The proposed DL model, designed based on the convolutional neural network architecture, combines a feature pyramid network (FPN) with texture encoding and deep supervision. We extracted, represent, and fused four-scale texture features to improve classification performance on high-risk local lesions. We also designed an auxiliary classification mechanism based on deep supervision to adjust the weight of each scale in FPN adaptively, enabling low-cost training of the whole model. Results: In the binary classification task detecting positive subjects with high-risk cervical lesions, our DL model achieved an 81.55% (95% CI, 72.70-88.51%) F1-score with 82.35% (95% CI, 69.13-91.60%) sensitivity and 81.48% (95% CI, 68.57-90.75%) specificity on the Renmin dataset, outperforming five experienced medical experts. It also achieved an 84.34% (95% CI, 74.71-91.39%) F1-score with 87.50% (95% CI, 73.20-95.81%) sensitivity and 90.59% (95% CI, 82.29-95.85%) specificity on the Huaxi dataset, comparable to the overall level of the best investigator. Moreover, our DL model provides visual diagnostic evidence of histomorphological and texture features learned in OCT images to assist gynecologists in making clinical decisions quickly. Conclusions: Our DL model holds great promise to be used in cervical lesion screening with OCT efficiently and effectively.

Multi-scale convolution enhanced transformer for multivariate long-term time series forecasting.

In data analysis and forecasting, particularly for multivariate long-term time series, challenges persist. The Transformer model in deep learning methods has shown significant potential in time series forecasting. The Transformer model's dot-product attention mechanism, however, due to its quadratic computational complexity, impairs training and forecasting efficiency. In addition, the Transformer architecture has limitations in modeling local features and dealing with multivariate cross-dimensional dependency relationship. In this article, a Multi-Scale Convolution Enhanced Transformer model (MSCformer) is proposed for multivariate long-term time series forecasting. As an alternative to modeling the time series in its entirety, a segmentation strategy is designed to convert the input original series into segmented forms with different lengths, then process time series segments using a new constructed multi-Dependency Aggregation module. This multi-Scale segmentation approach reduces the computational complexity of the attention mechanism part in subsequent models, and for each segment of length corresponds to a specific time scale, it also ensures that each segment retains the semantic information of the data sequence level, thereby comprehensively utilizing the multi-scale information of the data while more accurately capturing the real dependency of the time series. The Multi-Dependence Aggregate module captures both cross-temporal and cross-dimensional dependencies of multivariate long-term time series and compensates for local dependencies within the segments thereby captures local series features comprehensively and addressing the issue of insufficient information utilization. MSCformer synthesizes dependency information extracted from various temporal segments at different scales and reconstructs future series using linear layers. MSCformer exhibits higher forecasting accuracy, outperforming existing methods in multiple domains including energy, transportation, weather, electricity, disease and finance.

MFAM-AD: an anomaly detection model for multivariate time series using attention mechanism to fuse multi-scale features.

Multivariate time series anomaly detection has garnered significant attention in fields such as IT operations, finance, medicine, and industry. However, a key challenge lies in the fact that anomaly patterns often exhibit multi-scale temporal variations, which existing detection models often fail to capture effectively. This limitation significantly impacts detection accuracy. To address this issue, we propose the MFAM-AD model, which combines the strengths of convolutional neural networks (CNNs) and bi-directional long short-term memory (Bi-LSTM). The MFAM-AD model is designed to enhance anomaly detection accuracy by seamlessly integrating temporal dependencies and multi-scale spatial features. Specifically, it utilizes parallel convolutional layers to extract features across different scales, employing an attention mechanism for optimal feature fusion. Additionally, Bi-LSTM is leveraged to capture time-dependent information, reconstruct the time series and enable accurate anomaly detection based on reconstruction errors. In contrast to existing algorithms that struggle with inadequate feature fusion or are confined to single-scale feature analysis, MFAM-AD effectively addresses the unique challenges of multivariate time series anomaly detection. Experimental results on five publicly available datasets demonstrate the superiority of the proposed model. Specifically, on the datasets SMAP, MSL, and SMD1-1, our MFAM-AD model has the second-highest F1 score after the current state-of-the-art DCdetector model. On the datasets NIPS-TS-SWAN and NIPS-TS-GECCO, the F1 scores of MAFM-AD are 0.046 (6.2%) and 0.09 (21.3%) higher than those of DCdetector, respectively(the value ranges from 0 to 1). These findings validate the MFAMAD model's efficacy in multivariate time series anomaly detection, highlighting its potential in various real-world applications.

Relative position matrix and multi-scale feature fusion for writer-independent online signature verification.

Online signature verification (OSV) is widely used in finance, law and other fields, and is one of the important research projects on biological characteristics. However, its data set has a small scale and has high requirements for generalization of certification models. Therefore, how to overcome these problems is of great value to improve the practicality and security of online handwriting signature technology. We propose a writer-independent online handwritten signature verification method, which adopts the relative position matrix method to convert the traditional temporal features into images for processing. This method enriched the features of the signatures, serving the purpose of data augmentation. Then two-dimensional multi-scale feature fusion based Siamese neural network (2D-MFFnet) is built for representing and learning the importance of each channel adaptively combined with the attention mechanism. Finally, a temporal convolutional network is designed to construct the classifier. The results illustrate that compared with traditional time series models, the algorithm has reduced the equal error rate by at least 2.52 % on the open datasets MCYT-100 and SVC2004 task2.