Search Results (2,032)

This study investigates the application of deep learning (DL) techniques for predicting temperature fields in the SM1 swirl-stabilized turbulent non-premixed flame. Two distinct DL approaches were developed using a comprehensive CFD database generated via the steady laminar flamelet model coupled with the SST k-ω turbulence model. The first approach employs a fully connected dense neural network to directly map scalar input parameters—fuel velocity, swirl ratio, and equivalence ratio—to high-resolution temperature contour images. In addition, a comparison was made with different deep learning networks, namely Res-Net, EfficientNetB0, and Inception Net V3, to better understand the performance of the model. In the first approach, the results of the Inception V3 model and the developed Dense Model were found to be better than Res-Net and Efficient Net. At the same time, file sizes and usability were examined. The second framework employs a U-Net-based convolutional neural network enhanced by an RGB Fusion preprocessing technique, which integrates multiple scalar fields from non-reacting (cold flow) conditions into composite images, significantly improving spatial feature extraction. The training and validation processes for both models were conducted using 80% of the CFD data for training and 20% for testing, which helped assess their ability to generalize new input conditions. In the secondary approach, similar to the first approach, studies were conducted with different deep learning models, namely Res-Net, Efficient Net, and Inception Net, to evaluate model performance. The U-Net model, which is well developed, stands out with its low error and small file size. The dense network is appropriate for direct parametric analyses, while the image-based U-Net model provides a rapid and scalable option to utilize the cold flow CFD images. This framework can be further refined in future research to estimate more flow factors and tested against experimental measurements for enhanced applicability. Full article

Infrared small-target detection remains a critical challenge in military reconnaissance, environmental monitoring, forest-fire prevention, and search-and-rescue operations, owing to the targets’ extremely small size, sparse texture, low signal-to-noise ratio, and complex background interference. Traditional convolutional neural networks (CNNs) struggle to detect such weak, low-contrast objects due to their limited receptive fields and insufficient feature extraction capabilities. To overcome these limitations, we propose a Multi-Scale Edge-Aware Convolution (MEAC) module that enhances feature representation for small infrared targets without increasing parameter count or computational cost. Specifically, MEAC fuses (1) original local features, (2) multi-scale context captured via dilated convolutions, and (3) high-contrast edge cues derived from differential Gaussian filters. After fusing these branches, channel and spatial attention mechanisms are applied to adaptively emphasize critical regions, further improving feature discrimination. The MEAC module is fully compatible with standard convolutional layers and can be seamlessly embedded into various network architectures. Extensive experiments on three public infrared small-target datasets (SIRSTD-UAVB, IRSTDv1, and IRSTD-1K) demonstrate that networks augmented with MEAC significantly outperform baseline models using standard convolutions. When compared to eleven mainstream convolution modules (ACmix, AKConv, DRConv, DSConv, LSKConv, MixConv, PConv, ODConv, GConv, and Involution), our method consistently achieves the highest detection accuracy and robustness. Experiments conducted across multiple versions, including YOLOv10, YOLOv11, and YOLOv12, as well as various network levels, demonstrate that the MEAC module achieves stable improvements in performance metrics while slightly increasing computational and parameter complexity. These results validate the MEAC module’s significant advantages in enhancing the detection of small and weak objects and suppressing interference from complex backgrounds. These results validate MEAC’s effectiveness in enhancing weak small-target detection and suppressing complex background noise, highlighting its strong generalization ability and practical application potential. Full article

Structural health monitoring in resource-constrained environments demands crack segmentation models that match the accuracy of heavyweight convolutional networks while conforming to the power, memory, and latency limits of watt-level edge devices. This study presents a lightweight dual-attention network, which is a four-stage U-Net compressed to one-quarter of the channel depth and augmented—exclusively at the deepest layer—with a compact dual-attention block that couples channel excitation with spatial self-attention. The added mechanism increases computation by only 19%, limits the weight budget to 7.4 MB, and remains fully compatible with post-training INT8 quantization. On a pixel-labelled concrete crack benchmark, the proposed network achieves an intersection over union of 0.827 and an F1 score of 0.905, thus outperforming CrackTree, Hybrid 2020, MobileNetV3, and ESPNetv2. While refined weight initialization and Dice-augmented loss provide slight improvements, ablation experiments show that the dual-attention module is the main factor influencing accuracy. With 110 frames per second on a 10 W Jetson Nano and 220 frames per second on a 5 W Coral TPU achieved without observable accuracy loss, hardware-in-the-loop tests validate real-time viability. Thus, the proposed network offers cutting-edge crack segmentation at the kiloflop scale, thus facilitating ongoing, on-device civil infrastructure inspection. Full article

Wear state prediction based on oil monitoring technology enables the early identification of potential wear and failure risks of friction pairs, facilitating optimized equipment maintenance and extended service life. However, the complexity of lubricating oil monitoring data often poses challenges in extracting discriminative features, limiting the accuracy of wear state prediction. To address this, a CNN–LSTM–Attention network is specially constructed for predicting wear state, which hierarchically integrates convolutional neural networks (CNNs) for spatial feature extraction, long short-term memory (LSTM) networks for temporal dynamics modeling, and self-attention mechanisms for adaptive feature refinement. The proposed architecture implements a three-stage computational pipeline. Initially, the CNN performs hierarchical extraction of localized patterns from multi-sensor tribological signals. Subsequently, the self-attention mechanism conducts adaptive recalibration of feature saliency, prioritizing diagnostically critical feature channels. Ultimately, bidirectional LSTM establishes cross-cyclic temporal dependencies, enabling cascaded fully connected layers with Gaussian activation to generate probabilistic wear state estimations. Experimental results demonstrate that the proposed model not only achieves superior predictive accuracy but also exhibits robust stability, offering a reliable solution for condition monitoring and predictive maintenance in industrial applications. Full article

Translation initiation site (TIS) prediction in mRNA sequences constitutes an essential component of transcriptome annotation, playing a crucial role in deciphering gene expression and regulation mechanisms. Numerous computational methods have been proposed and achieved acceptable prediction accuracy. In our previous work, we developed NeuroTIS, a novel method for TIS prediction based on a hybrid dependency network combined with a deep learning framework that explicitly models label dependencies both within coding sequences (CDSs) and between CDSs and TISs. However, this method has limitations in fully exploiting the primary structural information within mRNA sequences. First, it only captures label dependency within three neighboring codon labels. Second, it neglects the heterogeneity of negative TISs originating from different reading frames, which exhibit distinct coding features in their vicinity. In this paper, under the framework of NeuroTIS, we propose its enhanced version, NeuroTIS+, which allows for more sophisticated codon label dependency modeling via temporal convolution and homogenous feature building through an adaptive grouping strategy. Tests on transcriptome-wide human and mouse datasets demonstrate that the proposed method yields excellent prediction performance, significantly surpassing the existing state-of-the-art methods. Full article

It is difficult for a single feature in wind power data to fully reflect the multifactor coupling relationship with wind power, while the forecast model hyperparameters rely on empirical settings, which affects the prediction accuracy. In order to effectively predict the continuous power in the future time period, an ultra-short-term prediction model of wind power based on fused features and an improved convolutional neural network (CNN) is proposed. Firstly, the historical power data are decomposed using dynamic modal decomposition (DMD) to extract their modal features. Then, considering the influence of meteorological factors on power prediction, the historical meteorological data in the sample data are extracted using kernel principal component analysis (KPCA). Finally, the decomposed power modal and the extracted meteorological components are reconstructed into multivariate time-series features; the snow ablation optimisation algorithm (SAO) is used to optimise the convolutional neural network (CNN) for wind power prediction. The results show that the root-mean-square error of the prediction result is 31.9% lower than that of the undecomposed one after using DMD decomposition; for the prediction of the power of two different wind farms, the root-mean-square error of the improved CNN model is reduced by 39.8% and 30.6%, respectively, compared with that of the original model, which shows that the proposed model has a better prediction performance. Full article

To address the issues of high model complexity and weak noise resistance in convolutional neural networks for bearing fault diagnosis, this paper proposes a novel lightweight multi-angle feature fusion convolutional neural network (LMAFCNN). First, the original signal was preprocessed using a wide-kernel convolutional layer to achieve data dimensionality reduction and feature channel expansion. Second, a lightweight multi-angle feature fusion module was designed as the core feature extraction unit. The main branch fused multidimensional features through pointwise convolution and large-kernel channel-wise expansion convolution, whereas the auxiliary branch introduced an efficient channel attention (ECA) mechanism to achieve channel-adaptive weighting. Feature enhancement was achieved through the addition of branches. Finally, global average pooling and fully connected layers were used to complete end-to-end fault diagnosis. The experimental results showed that the proposed method achieved an accuracy of 99.5% on the Paderborn University (PU) artificial damage dataset, with a computational complexity of only 14.8 million floating-point operations (MFLOPs) and 55.2 K parameters. Compared with existing mainstream methods, the proposed method significantly reduces model complexity while maintaining high accuracy, demonstrating excellent diagnostic performance and application potential. Full article

Reliable aerodynamic performance estimation is essential for both preliminary design and optimization in various aeronautical applications. In this study, a hybrid deep learning model is proposed, combining convolutional neural networks (CNNs) and operating directly on raw airfoil geometry, with parallel branches of fully connected deep neural networks (DNNs) that process operational parameters and engineered features. The model is trained on an extensive database of NACA four-digit airfoils, covering angles of attack ranging from −5° to 14° and ten Reynolds numbers increasing in steps of 500,000 from 500,000 up to 5,000,000. As a novel contribution, this work investigates the impact of random seed initialization on model accuracy and reproducibility and introduces a seed-based ensemble strategy to enhance generalization. The best-performing single-seed model tested (seed 0) achieves a mean absolute percentage error (MAPE) of 1.1% with an ${\mathrm{R}}^2$ of 0.9998 for the lift coefficient prediction and 0.57% with an ${\mathrm{R}}^2$ of 0.9954 for the drag coefficient prediction. In comparison, the best ensemble model tested (seeds 610, 987, and 75025) achieves a lift coefficient MAPE of 1.43%, corresponding to ${\mathrm{R}}^2$ 0.9999, and a drag coefficient MAPE of 1.19%, corresponding to ${\mathrm{R}}^2$ = 0.9968. All the tested seed dependencies in this paper (ten single seeds and five ensembles) demonstrate an overall ${\mathrm{R}}^2$ greater than 0.97, which reflects the model architecture’s strong foundation. The novelty of this study lies in the demonstration that the same machine learning model, trained on identical data and architecture, can exhibit up to 250% variation in prediction error solely due to differences in random seed selection. This finding highlights the often-overlooked impact of seed initialization on model performance and highlights the necessity of treating seed choice as an active design parameter in ML aerodynamic predictions. Full article

Rain streaks degrade the quality of optical images and seriously affect the effectiveness of subsequent vision-based algorithms. Although the applications of a convolutional neural network (CNN) and self-attention mechanism (SA) in single image deraining have shown great success, there are still unresolved issues regarding the deraining performance and the large computational load. The work in this paper fully coordinates and utilizes the advantages between CNN and SA and proposes a hybrid content-aware deraining network (CAD) to reduce complexity and generate high-quality results. Specifically, we construct the CADBlock, including the content-aware convolution and attention mixer module (CAMM) and the multi-scale double-gated feed-forward module (MDFM). In CAMM, the attention mechanism is used for intricate windows to generate abundant features and simple convolution is used for plain windows to reduce computational costs. In MDFM, multi-scale spatial features are double-gated fused to preserve local detail features and enhance image restoration capabilities. Furthermore, a four-token contextual attention module (FTCA) is introduced to explore the content information among neighbor keys to improve the representation ability. Both qualitative and quantitative validations on synthetic and real-world rain images demonstrate that the proposed CAD can achieve a competitive deraining performance. Full article

As global urbanization accelerates, communities have emerged as key areas where social conflicts and public safety risks clash. Traditional crime prevention models experience difficulties handling dynamic crime hotspots due to data lags and poor spatiotemporal resolution. Therefore, this study proposes a hybrid model combining Informer and Spatiotemporal Graph Convolutional Network (ST-GCN) to achieve precise crime prediction at the community level. By employing a community topology and incorporating historical crime, weather, and holiday data, ST-GCN captures spatiotemporal crime trends, while Informer identifies temporal dependencies. Moreover, the model leverages a fully connected layer to map features to predicted latitudes. The experimental results from 320,000 crime records from 22 police districts in Chicago, IL, USA, from 2015 to 2020 show that our model outperforms traditional and deep learning models in predicting assaults, robberies, property damage, and thefts. Specifically, the mean average error (MAE) is 0.73 for assaults, 1.36 for theft, 1.03 for robbery, and 1.05 for criminal damage. In addition, anomalous event fluctuations are effectively captured. The results indicate that our model furthers data-driven public safety governance through spatiotemporal dependency integration and long-sequence modeling, facilitating dynamic crime hotspot prediction and resource allocation optimization. Future research should integrate multisource socioeconomic data to further enhance model adaptability and cross-regional generalization capabilities. Full article

Field-based landslide mapping is a crucial task for geo-hydrological risk assessment but is often limited by the lack of integrated tools to capture accurate spatial and temporal data. This research investigates a Direct Numerical Cartography (DNC) system’s ability to capture both spatial and temporal landslide features during fieldwork. DNC enables fully digital surveys, minimizing errors and delivering real-time, spatially accurate data to experts on site. We tested an integrated approach combining the Ground Operative System for GIS Input Remote-data Acquisition (GOGIRA) with the Expert-based Multitemporal AI Detector (ExMAD). GOGIRA is a low-cost system for efficient georeferenced data collection, while ExMAD uses AI and multitemporal Sentinel-2 imagery to detect landslide triggering times. Upgrades to GOGIRA’s hardware and algorithms were carried out to improve its mapping accuracy. Field tests in Western Greece compared data to 64 expert-confirmed landslides, with the Range-R device showing a mean spatial error of 50 m, outperforming the tripod-based UGO device at 82 m. Operational factors like line-of-sight obstructions and terrain complexity affected accuracy. ExMAD applied a pre-trained U-Net convolutional neural network for automated temporal trend detection of landslide events. The combined DNC and AI-assisted remote sensing approach enhances landslide inventory precision and consistency while maintaining expert oversight, offering a scalable solution for landslide monitoring. Full article

As the scale of distribution networks expand, existing defect identification methods face numerous challenges, including limitations in single-modal feature identification, insufficient cross-modal information fusion, and the lack of a multi-stage feedback mechanism. To address these issues, we first propose a joint perception of image and voiceprint features based on bidirectional coupled attention, which enhances deep interaction across modalities and overcomes the shortcomings of traditional methods in cross-modal fusion. Secondly, a defect identification and diagnosis method of distribution network electrical equipment based on two-stage convolutional neural networks (CNN) is introduced, which makes the network pay more attention to typical and frequent defects, and enhances defect diagnosis accuracy and robustness. The proposed algorithm is compared with two baseline algorithms. Baseline 1 is a long short term memory (LSTM)-based algorithm that performs separate feature extraction and processing for image and voiceprint signals without coupling the features of the two modalities, and Baseline 2 is a traditional CNN algorithm that uses classical convolutional layers for feature learning and classification through pooling and fully connected layers. Compared with two baselines, simulation results demonstrate that the proposed method improves accuracy by 12.1% and 33.7%, recall by 12.5% and 33.1%, and diagnosis efficiency by 22.92% and 60.42%. Full article

To address the limitations in time–frequency feature representation of shearer arm gear faults and the issues of parameter redundancy and low training efficiency in standard convolutional neural networks (CNNs), this study proposes a diagnostic method based on an improved S-transform and a Depthwise Separable Convolutional Neural Network (DSCNN). First, the improved S-transform is employed to perform time–frequency analysis on the vibration signals, converting the original one-dimensional signals into two-dimensional time–frequency images to fully preserve the fault characteristics of the gear. Then, a neural network model combining standard convolution and depthwise separable convolution is constructed for fault identification. The experimental dataset includes five gear conditions: tooth deficiency, tooth breakage, tooth wear, tooth crack, and normal. The performance of various frequency-domain and time-frequency methods—Wavelet Transform, Fourier Transform, S-transform, and Gramian Angular Field (GAF)—is compared using the same network model. Furthermore, Grad-CAM is applied to visualize the responses of key convolutional layers, highlighting the regions of interest related to gear fault features. Finally, four typical CNN architectures are analyzed and compared: Deep Convolutional Neural Network (DCNN), InceptionV3, Residual Network (ResNet), and Pyramid Convolutional Neural Network (PCNN). Experimental results demonstrate that frequency–domain representations consistently outperform raw time-domain signals in fault diagnosis tasks. Grad-CAM effectively verifies the model’s accurate focus on critical fault features. Moreover, the proposed method achieves high classification accuracy while reducing both training time and the number of model parameters. Full article

Single-cell Assay for Transposase-Accessible Chromatin using sequencing (scATAC-seq) technology enables single-cell resolution analysis of chromatin accessibility, offering critical insights into gene regulation, epigenetic heterogeneity, and cellular differentiation across various biological contexts. However, existing cell annotation methods face notable limitations. Cross-omics approaches, which rely on single-cell RNA sequencing (scRNA-seq) as a reference, often struggle with data alignment due to fundamental differences between transcriptional and chromatin accessibility modalities. Meanwhile, intra-omics methods, which rely solely on scATAC-seq data, are frequently affected by batch effects and fail to fully utilize genomic sequence information for accurate annotation. To address these challenges, we propose scAttG, a novel deep learning framework that integrates graph attention networks (GATs) and convolutional neural networks (CNNs) to capture both chromatin accessibility signals and genomic sequence features. By utilizing the nucleotide sequences corresponding to scATAC-seq peaks, scAttG enhances both the robustness and accuracy of cell-type annotation. Experimental results across multiple scATAC-seq datasets suggest that scAttG generally performs favorably compared to existing methods, showing competitive performance in single-cell chromatin accessibility-based cell-type annotation. Full article

Deep learning has achieved remarkable success in hyperspectral image (HSI) classification, attributed to its powerful feature extraction capabilities. However, existing methods face several challenges: Convolutional Neural Networks (CNNs) are limited in modeling long-range spectral dependencies because of their limited receptive fields; Transformers are constrained by their quadratic computational complexity; and Mamba-based methods fail to fully exploit spatial–spectral interactions when handling high-dimensional HSI data. To address these limitations, we propose MRFP-Mamba, a novel Multi-Receptive-Field Parallel Mamba architecture that integrates hierarchical spatial feature extraction with efficient modeling of spatial–spectral dependencies. The proposed MRFP-Mamba introduces two key innovation modules: (1) A multi-receptive-field convolutional module employing parallel $1\times 1$, $3\times 3$, $5\times 5$, and $7\times 7$ kernels to capture fine-to-coarse spatial features, thereby improving discriminability for multi-scale objects; and (2) a parameter-optimized Vision Mamba branch that models global spatial–spectral relationships through structured state space mechanisms. Experimental results demonstrate that the proposed MRFP-Mamba consistently surpasses existing CNN-, Transformer-, and state space model (SSM)-based approaches across four widely used hyperspectral image (HSI) benchmark datasets: PaviaU, Indian Pines, Houston 2013, and WHU-Hi-LongKou. Compared with MambaHSI, our MRFP-Mamba achieves improvements in Overall Accuracy (OA) by 0.69%, 0.30%, 0.40%, and 0.97%, respectively, thereby validating its superior classification capability and robustness. Full article