Multiscale Strategy Research Articles

Robust multi-stage progressive autoencoder for hyperspectral anomaly detection

Recently, Autoencoders (AEs) have demonstrated remarkable performance in the field of hyperspectral anomaly detection, owing to their powerful capability in handling high-dimensional data. However, they often overlook the inherent global distribution characteristics and long-range dependencies in hyperspectral images (HSI). This oversight makes it challenging to accurately characterize and describe boundaries between different backgrounds and anomalies in complex HSI, thereby affecting detection accuracy. To address this issue, a robust multi-stage progressive autoencoder for hyperspectral anomaly detection (RMSAD) is proposed. Initially, a progressive multi-stage learning framework based on convolutional autoencoders is employed. This framework incrementally reveals and integrates deep contextual features along with their long-range dependencies in HSI, aiming to accurately characterize the background and anomalies. Subsequently, an innovative multi-scale fusion strategy is introduced at the intersections of each stage, reinforcing the learning and representation of background and global spatial details across multiple stages. Finally, by collectively extracting abnormal spatial information across stages, effectively reducing the tendency of autoencoders to reconstruct anomalies. This ensures the efficient restoration and replication of global textural details in HSI. The experimental results on the six HSI datasets demonstrate that the proposed RMSAD is superior to other state-of-the-art methods.

Differentiable design of freeform diffractive optical elements for beam shaping by representing phase distribution using multi-level B-splines

The generation of a specific laser beam profile on the work surface is key to various laser beam shaping tasks, relying heavily on diffractive optical elements (DOEs). Most beam-shaping DOEs are designed using iterative Fourier transform algorithms (IFTAs), which generally have slow convergence and prone to stagnate at local minima. Moreover, the microreliefs generated by IFTAs tend to be irregular, complicating manufacturing and causing uncontrolled scattering of light. We propose a differentiable DOE design method that applies a phase-smoothness constraint using multi-level B-splines. A multi-scale gradient-descent optimization strategy, naturally linked with the multi-level B-splines, is employed to robustly determine the optimized phase distribution that is fully continuous. This, in turn, can lead to more regular DOE microreliefs, which can simplify the fabrication process and be less sensitive to changes in wavelength and working distance. Furthermore, our method can also design a fully continuous freeform lens, distinguished from most freeform lens design approaches by its foundation in physical optics rather than geometrical optics. Simulation and experimental results of several design tasks demonstrate the effectiveness of the proposed method.

Rational MOF Membrane Design for Gas Detection in Complex Environments.

Metal-organic frameworks (MOFs) hold significant promise in the realm of gas sensing. However, current understanding of their sensing mechanisms remains limited. Furthermore, the large-scale fabrication of MOFs is hampered by their inadequate mechanical properties. These two challenges contribute to the sluggish development of MOF-based gas-sensing materials. In this review, the selection of metal ions and organic ligands for designing MOFs is first presented, deepening the understanding of the interactions between different metal ions/organic ligands and target gases. Subsequently, the typical interfacial synthesis strategies (gas-solid, gas-liquid, solid-liquid interfaces) are provided, highlighting the potential for constructing MOF membranes on superhydrophobic and/or superhydrophilic substrates. Then, a multi-scale structure design strategies is proposed, including multi-dimensional membrane design and heterogeneous membrane design, to improve sensing performance through enhanced interfacial mass transfer and specific gas sieving. This strategy is anticipated to augment the task-specific capabilities of MOF-based materials in complex environments. Finally, several key future research directions are outlined with the aim not only to further investigate the underlying sensing principles of MOF membranes but also to achieve efficient detection of target gases amidst interfering gases and elevated moisture levels.

Post-earthquake inspection of high-speed railway viaducts with multi-scale task interaction deep learning strategy

Automated computer vision-based inspections of railway infrastructures, such as component type, damaged status, and location, have been investigated actively by resorting to task-specific deep learning models. However, task-specific models that fulfill these separate inspection tasks encountered bottlenecks in improving inference accuracy, and bring huge computational costs. Multi-task deep learning, which can fulfill these inspection tasks concurrently, has yet to be fully investigated in the context of structural inspection. In this study, a multi-scale task interaction deep learning strategy is presented towards component recognition, damage segmentation, and depth estimation for a comprehensive post-earthquake inspection of high-speed railway viaducts. Three modules for multi-scale task interaction were proposed to modify the multi-task deep neural network, taking full advantage of task commonalities at multiple scales. The proposed method was validated with a large-scale image dataset of high-speed railway viaducts. Component recognition and depth estimation were incorporated to implement multi-task learning since they have higher pattern affinities at multiple scales. Results reported that mean Intersection over Unions of testing samples of component and damage tasks were 91.2% and 72.1%, RMSE of depth estimation was 1.54 m. Compared with single-task cases, training time, inference duration, and FLOPs of the multi-task model were reduced by 23%, 30%, 27% respectively. Results showed improvement in both inference accuracy and training efficiency, substantiating the superiority of the proposed strategy.

Engineered Protein Fibers with Reinforced Mechanical Properties Via β-Sheet High-Order Assembly.

Protein fibers are ideal alternatives to synthetic polymers due to their unique mechanical properties, biocompatibility, and sustainability. However, engineering biomimetic protein fibers with high mechanical properties remains challenging, particularly in mimicking the high molecular weight of natural proteins and regulating their complex hierarchical structures. Here, a modular design and multi-scale assembly strategy is developed to manufacture robust protein fibers using low- or medium-molecular-weight proteins. The distinct functional and structural properties of flexible, rigid, and cross-linked domains in modular proteins are skillfully harnessed. By regulating the ratio of rigid to flexible domains, the formation of high-order β-sheet crystals aligned along the fiber axis is promoted, enhancing both strength and toughness. Furthermore, the dynamic imine cross-linking network, formed by the aldehyde-amine condensation reaction of the cross-linked domains, further reinforces the protein fibers. Remarkably, fibers spun from modular proteins significantly smaller than natural spidroin exhibit outstanding mechanical properties, surpassing those of protein fibers with same or even higher molecular weights. This strategy offers a promising pathway for fabricating protein fibers suitable for diverse applications.

Waste classification strategy based on multi-scale feature fusion for intelligent waste recycling in office buildings

Waste classification is an important measure to protect the environment. Existing waste classification methods mainly focus on scientific research, but lack attention to the challenges of waste classification in actual scenarios. For example, wastes with similar contours, similar textures, or contaminated appearance are difficult to be classified in actual scenarios. To address these issues, this paper proposes an innovative multi-scale feature fusion strategy (MFFS) to improve the classification accuracy of these wastes. MFFS combines local fine-grained features with global coarse-grained features to improve the feature expression ability of waste. However, how to effectively fuse these two features is a key challenge. This paper proposes a dual-scale feature fusion strategy, first fusing fine-grained features in the first dimension, then fusing coarse-grained features in the second dimension, and introducing spatial features to further enhance feature expression capabilities. In order to reduce the interference of background information, the model in this paper models global relationships based on convolutional features. The MFFS strategy achieved a classification accuracy of 95.5% on the self-built dataset and 94.1% on the public dataset TrashNet. The number of parameters of our model is reduced by 57.2% compared with the classic VGG16 and by 34.2% compared with the Vision Transformer. In addition, we designed an intelligent waste sorting device and deployed the MFFS model on the device to implement the application. Experiments show that our model has ideal accuracy and stability and can be promoted and applied.

Ferroelectric tungsten bronze-based ceramics with high-energy storage performance via weakly coupled relaxor design and grain boundary optimization

A multiscale regulation strategy has been demonstrated for synthetic energy storage enhancement in a tetragonal tungsten bronze structure ferroelectric. Grain refining and second-phase precipitation (perovskite phase) are introduced in the BaSrTiNb2-xTaxO9 ceramics by regulating the composition and sintering process. Disordered polarization and distribution, chemical inhomogeneity, and insulating boundary layers are achieved to provide the fundamental structural origin of the relaxation characteristic, high breakdown strength, and superior energy storage performance. Thus, an ultrahigh energy storage density of 12.2 J cm−3 with an low energy consumption was achieved at an electric field of 950 kV cm−1. This is the highest known energy storage performance in tetragonal tungsten bronze-based ferroelectric. Notably, this ceramic shows remarkable stability over frequency, temperature, and cycling electric fields. This work brings new material candidates and structure design for developing of energy storage capacitors apart from the predominant perovskite ferroelectric ceramics.

Multiscale molecular modeling of chromatin with MultiMM: From nucleosomes to the whole genome

Motivation: We present a user-friendly 3D chromatin simulation model for the human genome based on OpenMM, addressing the challenges posed by existing models with use-specific implementations. Our approach employs a multi-scale energy minimization strategy, capturing chromatin's hierarchical structure. Initiating with a Hilbert curve-based structure, users can input files specifying nucleosome positioning, loops, compartments, or subcompartments.Results: The model utilizes an energy minimization approach with a large choice of numerical integrators, providing the entire genome's structure within minutes. Output files include the generated structures for each chromosome, offering a versatile and accessible tool for chromatin simulation in bioinformatics studies. Furthermore, MultiMM is capable of producing nucleosome-resolution structures by making simplistic geometric assumptions about the structure and the density of nucleosomes on the DNA.Code availability: Open-source software and the manual are freely available on https://github.com/SFGLab/MultiMM or via pip https://pypi.org/project/MultiMM/.

MT-CrackNet:A multi-task deep learning framework for automatic in-situ fatigue micro-crack detection and quantification

Characterizing fatigue micro-cracks is crucial for understanding the mechanisms and behaviors of material damage. In-situ fatigue testing is an essential method for observing the evolution of fatigue micro-cracks; however, the process often requires significant time, making the measurement of micro-cracks a tedious task. This paper introduces a multi-task deep learning framework called MT-CrackNet, which enables automatic detection and quantification of in-situ fatigue micro-cracks. The framework is capable of recognizing or segmenting multiple tasks such as micro-cracks, text, and scales simultaneously, and its effectiveness is not limited by the magnification of in-situ images. By integrating attention mechanisms and multi-scale strategies, the model enhances its ability to handle long-range dependencies and preserve detail information, accurately identifying and measuring the length of micro-cracks. The effectiveness of the proposed MT-CrackNet is validated through three in-situ fatigue micro-crack propagation experiments.

A parallel and multi-scale probabilistic temporal convolutional neural networks for forecasting the key monitoring parameters of gas turbine

As a crucial equipment in the power industry, gas turbines need effective condition monitoring techniques to maintain safe and reliable operations. Fast yet accurate long-term forecasting of the key monitoring parameters of gas turbine is vital for achieving the overall effective condition monitoring. This however poses challenges in terms of both efficacy and efficiency for conventional time series models like recurrent neural networks (RNN), since they run in a sequential manner. To this end, this paper introduces a novel parallel probabilistic time series prediction model. Particularly, the proposed model leverages the parallelism of temporal convolutional neural network (TCN) and exploits the power of latent space in the Bayesian framework. Besides, a multi-scale feature extraction strategy based on dense connections and a lagged observation strategy are presented to improve the model performance. In the comparative study against state-of-the art competitors on four classical system identification benchmarks, the proposed model significantly reduces the training time, while consistently achieving highly accurate predictions and providing appropriate uncertainty quantification. Thereafter, the proposed model is adopted to build a systematic workflow for the fast yet accurate long-term forecasting of the temperature of blade channel of gas turbine. The comprehensive results again highlight the superiority of the proposed model in terms of both the prediction quality and the training efficiency.

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 areinvestigated. 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.

PMSFF: Improved Protein Binding Residues Prediction through Multi-Scale Sequence-Based Feature Fusion Strategy.

Proteins perform different biological functions through binding with various molecules which are mediated by a few key residues and accurate prediction of such protein binding residues (PBRs) is crucial for understanding cellular processes and for designing new drugs. Many computational prediction approaches have been proposed to identify PBRs with sequence-based features. However, these approaches face two main challenges: (1) these methods only concatenate residue feature vectors with a simple sliding window strategy, and (2) it is challenging to find a uniform sliding window size suitable for learning embeddings across different types of PBRs. In this study, we propose one novel framework that could apply multiple types of PBRs Prediciton task through Multi-scale Sequence-based Feature Fusion (PMSFF) strategy. Firstly, PMSFF employs a pre-trained language model named ProtT5, to encode amino acid residues in protein sequences. Then, it generates multi-scale residue embeddings by applying multi-size windows to capture effective neighboring residues and multi-size kernels to learn information across different scales. Additionally, the proposed model treats protein sequences as sentences, employing a bidirectional GRU to learn global context. We also collect benchmark datasets encompassing various PBRs types and evaluate our PMSFF approach to these datasets. Compared with state-of-the-art methods, PMSFF demonstrates superior performance on most PBRs prediction tasks.

Multiscale modelling strategy for predicting fatigue performance of welded joints

This study predicts the fatigue performance of welded joints through a multiscale modelling strategy accounting for material and structural inhomogeneities. An S-N curve and detailed fracture surfaces with distinct beach marks were first derived by uniaxial fatigue tests utilising the designed cruciform welded joints. Considering the intrinsic features of welded joints, a multiscale modelling strategy was proposed to integrate multiple factors, including microstructural variations, strength distributions within the heat-affected zone (HAZ), and the diversity of three-dimensional weld toe shapes. Significantly, a modelling strategy was presented for the first time to simulate the simultaneous initiation, growth, and coalescence of multiple cracks, and was validated against experimental evidence. The results indicate that the proposed strategy can accurately predict both the fatigue strength and the overall crack growth process. Additionally, comparative assessments of single-crack and multiple-crack modelling strategies revealed notably shorter predicted fatigue lives when considering crack coalescence. Overall, this work establishes a multiscale framework for assessing the fatigue performance of welded joints considering both microscopic and macroscopic factors, offering substantial practical implications for engineering applications.

Firefighting Robot Extinguishment Decision‐Making Based on Visual Guidance: A Novel Attention and Scale U‐Net Model and Genetic Algorithm

ABSTRACTAt present, rescue firefighting relies mainly on firefighting robots, and robots with perception and decision‐making functions are the key elements for achieving intelligent firefighting. However, traditional firefighting robots often lack the ability for autonomous perception and decision‐making when extinguishing multiple fire sources, leading to low rescue efficiency and increased risk for rescue personnel, especially when making firefighting decisions in extreme fire scenes, which poses a challenge. To effectively handle firefighting tasks and ensure operational efficiency, a robot firefighting decision‐making method based on drone visual guidance is proposed. First, we introduce a novel Attention and Scale U‐Net (ASUNet) model to accurately capture crucial target information, including fire location and size, in a fire scene. The ASUNet model adopts an effective multiscale fusion strategy and attention mechanism to enhance the model's performance. Subsequently, based on the results of the ASUNet model, through pixel segmentation clustering and a genetic optimization algorithm, we obtain the robot's firefighting decision results, thereby guiding the robot to carry out firefighting operations systematically. Finally, through numerical experiments, it is verified that the proposed ASUNet model is superior and effective, as the model can perceive important information in a fire scene and extract it well. The use of improved genetic optimization can further accelerate algorithm convergence. To our knowledge, this study is the first to use drone‐based monocular vision guidance for firefighting decision‐making, providing significant engineering value.

Real-Time ConvNext-Based U-Net with Feature Infusion for Egg Microcrack Detection

Real-time automatic detection of microcracks in eggs is crucial for ensuring egg quality and safety, yet rapid detection of micron-scale cracks remains challenging. This study introduces a real-time ConvNext-Based U-Net model with Feature Infusion (CBU-FI Net) for egg microcrack detection. Leveraging edge features and spatial continuity of cracks, we incorporate an edge feature infusion module in the encoder and design a multi-scale feature aggregation strategy in the decoder to enhance the extraction of both local details and global semantic information. By introducing large convolution kernels and depth-wise separable convolution from ConvNext, the model significantly reduces network parameters compared to the original U-Net. Additionally, a composite loss function is devised to address class imbalance issues. Experimental results on a dataset comprising over 3400 graded egg microcrack image patches demonstrate that CBU-FI Net achieves a reduction in parameters to one-third the amount in the original U-Net, with an inference speed of 21 ms per image (1 million pixels). The model achieves a Crack-IoU of 65.51% for microcracks smaller than 20 μm and a Crack-IoU and MIoU of 60.76% and 80.22%, respectively, for even smaller cracks (less than 5 μm), achieving high-precision, real-time detection of egg microcracks. Furthermore, on the publicly benchmarked CrackSeg9k dataset, CBU-FI Net achieves an inference speed of 4 ms for 400 × 400 resolution images, with an MIoU of 81.38%, proving the proposed method’s robustness and generalization capability across various cracks and complex backgrounds.

Tensile performance prediction of CFRPs with voids using multiscale analysis and neural networks

The overall work provides a novel multiscale modeling strategy applicable to the mechanical response prediction of arbitrarily laid porous laminates under uniaxial tension by establishing representative defective volumes (RDVs) for the inter-ply and intra-ply regions. A representative defective volume modeling method based on void characterization results using X-ray computed tomography for the interlaminar region was proposed. Elastic properties, tensile strength, and fracture energies of porous and poreless microscale models were obtained and used in the macroscale model. Defective properties were assigned to partial elements in the macroscale model according to the preset distribution mode. The Extended Finite Element Method was taken to simulate the failure process. Tensile properties of porous laminates with stacking sequences of [90]8, [90/0/90/0]2S, and [45/90/-45/0]2S were experimentally tested and numerically simulated. The predicted tensile properties and crack propagation behavior coincide well with the test results. Besides, the effects of micro-voids and macroscopic void distribution were discussed in detail. Further, a back-propagation neural network was used to achieve a fast prediction of the tensile properties and save computational costs. The computation time (<1 s) greatly improved compared with multiscale modeling (>38 h). This work provides an efficient tool for composite performance prediction, design, and optimization.

Multi-scale matching and simulating flows of ecosystem service supply and demand in the Wuhan metropolitan area, China

Metropolitan areas face the challenge of balancing increasing social needs against limited natural resources. The multi-scale evaluation of ecosystem services (ES) supply, demand and flows is crucial but under explored. Four prominent ES, food provision (FP), carbon sequestration (CS), soil conservation (SC), and outdoor recreation (OR) in Wuhan metropolitan area (WMA) were selected as cases. ES supply and demand were assessed by InVEST model, geographic information system (GIS) spatial analysis, and statistical methods. A linear programming model was employed to reveal the characteristics of ES flows. Results showed the following: (1) the total ES supply capacity can meet the demand of the study area. At different scales, the supply patterns were fundamentally consistent, except for FP, which was significantly higher in Wuhan's suburban areas. In contrast, the other three ES exhibited high supply in the northeast and south WMA, encompassing the Dabie and Mufu Mountains. The high demand for all ES radiated from the city center, except for SC. (2) A significant scale effect of ES mismatch was found, and ecological deficit patches decreased with increasing research scale. The Mismatch was evident on finer scales but lessened at coarser ones. (3) ES flows, consisting of FP, CS, and OR, formed a flow pattern from the suburbs to the city center. Coarse-scale management integrating ecological flows may better maintain ES balance than fine-scale measures. These findings provide insights for multi-scale management strategies that not only contribute to localized ecological restoration, but also optimize strategies for ecological corridors and compensation mechanisms.

A long-term dissolved oxygen prediction model in aquaculture using transformer with a dynamic adaptive mechanism

The prediction of Dissolved Oxygen (DO) in oceans is crucial for the survival of marine life and the management of marine farms. However, due to the inherent instability and uncertainty of DO, prediction models frequently necessitate parameter adjustments, posing challenges to maintaining reliable predictive performance and generalization capabilities. Based on this, a long-term dissolved oxygen prediction model in aquaculture using Transformer with a dynamic adaptive mechanism is proposed in the paper. A multi-scale latent correlation decomposition strategy was devised, merging both time and frequency domain decompositions. This approach proficiently untangles time series exhibiting intricate non-linear cycles, thereby delving deeper into and furnishing a more comprehensive array of feature data. Furthermore, an uncertainty-aware attention mechanism was posited, ensuring accurate pinpointing of pivotal data amidst a highly volatile data environment, thus cementing prediction accuracy. Finally, the incorporation of an adaptive hyperparameter adjustment mechanism enables the model to recalibrate in line with feature insights, bolstering its generalization potential across marine farms in diverse regions. Empirical results show that the model proposed in this study can effectively accomplish high-precision long-term forecasting. When juxtaposed with other models under comparable conditions, its mean squared error significantly diminished by 40.2%, the mean absolute error contracted by 55.4%, and concurrently, the R2 value ascended by 7%. This research proffers a fresh vantage point for long-term marine chemical data predictions and extends robust technical bolstering for refining and augmenting aquacultural strategies.

Enhanced Energy Storage Properties of Highly Polarized BMT-Based Thin Films through the Multiscale Structure Synergistic Regulation Strategy.

For solving the trade-off relationship of the polarization and breakdown electric field, ferroelectric films with high polarization are playing a critical role in energy storage capacitor applications, especially at moderate/low electric fields. In this work, we propose a multiscale structure (including defect, domain, and grain structures) synergetic optimization strategy to optimize the polarization behavior and energy storage performances of BiMg0.5Ti0.5O3 (BMT) ferroelectric films by introducing Sr0.7La0.2TiO3 (SLT) without compromising the breakdown strength. At a moderate electric field of 2917 kV/cm, a high discharge density of 72.2 J/cm3 has been achieved in 0.9BMT-0.1SLT films, together with good frequency, thermal, and cycle stabilities for energy storage. Importantly, the phase difference Δφ is utilized to quantitatively evaluate the polarization switching mobility of the ferroelectric domain/PNRs at an external electric field stimulation. This research demonstrates that a multiscale structure optimization strategy could effectively regulate the energy storage performance, and ecofriendly BMT-based materials are promising candidates for next-generation energy storage capacitors, especially at moderate/low electric fields.

Model development using hybrid method for prediction of drug release from biomaterial matrix

A comprehensive multi-scale computational strategy was developed in this study based on mass transfer and machine learning for simulation of drug concentration distribution in a biomaterial matrix. The controlled release was modeled and validated via the hybrid model. Mass transfer equations along with kinetics models were solved numerically and the results were then used for machine learning models. We investigated the performance of three regression models, namely Decision Tree (DT), Random Forest (RF), and Extra Tree (ET) in predicting medicine concentration (C) based on r and z data. Hyper-parameter optimization is conducted using Glowworm Swarm Optimization (GSO). Results revealed high predictive accuracy across all models, with ET demonstrating superior performance, achieving a coefficient of determination value (R2) of 0.99854, an RMSE of 1.1446E-05, and a maximum error of 6.49087E-05. DT and RF also exhibit notable performance, with coefficients of determination equal to 0.99571 and 0.99655, respectively. These results highlight the effectiveness of ensemble tree-based methods in accurately predicting chemical concentrations, with Extra Tree (ET) Regression emerging as the most promising model for this specific dataset.