Convolutional Attention Mechanism Research Articles

Text Detection on Industrial Barrel Label with Convolutional Attention and Dual‐Branch Feature Network

Industrial barrel labels generally have low visual contrast, uneven lighting, and cluttered background, making it challenging to accurately locate text regions. This paper proposes a text detection network to solve the inaccurate localization problem based on DBNet. First, a convolutional attention mechanism is applied to the feature extraction network to get more valuable text feature maps. Then, a dual‐branch convolutional feature module is proposed in the feature pyramid to enrich contextual information. Besides, during the probability map generation stage, using a feature remodeling enhancement module to further distinguish text and text boundaries. This paper designs comparative experiments on ILTD, ICDAR2015 and MSRA‐TD500 datasets, achieve F‐measure of 92.3%, 86.0% and 84.1%, which are 2.2%, 2.3%, and 1.9% higher than DBNet, respectively. They demonstrate that our proposed method exhibits competitive performance and strong robustness. © 2024 Institute of Electrical Engineers of Japan and Wiley Periodicals LLC.

Three-dimensional dynamic gesture recognition method based on convolutional neural network

With the rapid advancement of virtual reality, dynamic gesture recognition technology has become an indispensable and critical technique for users to achieve human–computer interaction in virtual environments. The recognition of dynamic gestures is a challenging task due to the high degree of freedom and the influence of individual differences and the change of gesture space. To solve the problem of low recognition accuracy of existing networks, an improved dynamic gesture recognition algorithm based on ResNeXt architecture is proposed. The algorithm employs three-dimensional convolution techniques to effectively capture the spatiotemporal features intrinsic to dynamic gestures. Additionally, to enhance the model’s focus and improve its accuracy in identifying dynamic gestures, a lightweight convolutional attention mechanism is introduced. This mechanism not only augments the model’s precision but also facilitates faster convergence during the training phase. In order to further optimize the performance of the model, a deep attention submodule is added to the convolutional attention mechanism module to strengthen the network’s capability in temporal feature extraction. Empirical evaluations on EgoGesture and NvGesture datasets show that the accuracy of the proposed model in dynamic gesture recognition reaches 95.03% and 86.21%, respectively. When operating in RGB mode, the accuracy reached 93.49% and 80.22%, respectively. These results underscore the effectiveness of the proposed algorithm in recognizing dynamic gestures with high accuracy, showcasing its potential for applications in advanced human–computer interaction systems.

EEG Emotion Recognition Based on GADF and AMB‐CNN Model

ABSTRACTDeep learning has achieved better results in natural language processing, computer vision, and other fields. Nowadays, more deep learning algorithms have also been applied in brain‐based emotion recognition. In the studies on brain‐based emotion recognition, deep learning models typically use one‐dimensional time series as the input and cannot fully leverage the advantages of the models in image classification or recognition. To address this issue, based on the publicly available SEED and DEAP datasets, the Gramian angular difference field (GADF) method was proposed to construct two‐dimensional image representation datasets: SEED‐GADF and DEAP‐GADF datasets, in the paper. Additionally, a convolutional attention mechanism model (AMB‐CNN) was introduced and its classification performance was validated on SEED‐GADF and DEAP‐GADF datasets. AMB‐CNN achieved an average accuracy of 90.8%, a recall rate of 90%, and AUC of 96.86% on SEED‐GADF. On DEAP‐GADF, the average accuracy, recall rate, and AUC respectively reached 96.06%, 96.06%, and 98.58% in the valence dimension and 96.11%, 96.11%, and 98.73% in the arousal dimension. Finally, the comparison results with various algorithms and ablation experiments proved the superiority of the proposed model.

Robust and energy-efficient RPL optimization algorithm with scalable deep reinforcement learning for IIoT

The increasing complexity and quantity of the Industrial Internet of Things (IIoT) pose new challenges to the traditional routing protocol for low-power and lossy networks (RPL) in terms of dynamic management, data transmission reliability, and energy efficiency optimization. This paper proposes a scalable deep reinforcement learning (DRL) algorithm with a multi-attention actor double critic model for routing optimization (MADC) to meet the requirements of IIoT for efficient and intelligent routing decisions while improving data transmission reliability and energy efficiency. Specifically, MADC employs the centralized training and decentralized execution (CTDE) learning paradigm to decouple the model’s training and inference tasks, which reduces the difficulty and computational cost of model learning and improves the training efficiency. In addition, a lightweight actor network based on multi-scale convolutional attention mechanism is designed in MADC, which can provide intelligent and real-time decision-making capabilities for resource-constrained nodes with low computational and storage complexities. Moreover, a scalable critic network utilizing multiple attention mechanisms is proposed. It is not only suitable for dynamic and changing network environments but also can more comprehensively and accurately evaluate local observation states, providing more accurate and efficient guidance for model optimization. Furthermore, MADC incorporates a double critic network architecture to mitigate potential overestimation issues during training, thereby ensuring the model’s robustness and reliability. Simulation results demonstrate that MADC outperforms existing RPL optimization algorithms in terms of energy efficiency, data transmission reliability, and adaptability.

High-resolution weld defect detection with RSU-MLP and dynamic kernel supervision

The challenges posed by high-resolution X-ray images of weld defects—including fuzzy features, multi-scale variability, small targets, and sample imbalance—create significant obstacles for accurate defect localization. Traditional deep learning methods typically emphasize global image characteristics, often neglecting local and multi-scale information, which leads to the inaccurate detection of small defects. To address this issue, we propose a novel high-resolution weld defect detection method called RSU-MLP. This method combines the RSU-MLP network with multi-scale feature extraction technology to effectively capture both local and global image features. Additionally, we design a dynamic kernel adaptive segmentation head based on a hybrid extended convolutional attention mechanism, which adaptively adjusts the receptive field size of the convolutional kernel, thereby enhancing detection capabilities for defects of varying scales. Furthermore, an in-depth supervision mechanism is introduced to monitor and learn defects at different levels of the network, leading to improved detection performance. Experimental results on real high-resolution welding datasets demonstrate that the proposed method achieves promising results in welding defect detection. Compared to traditional approaches, this method provides more accurate detection and identification of various welding defects, achieving DICE and Jaccard scores of 75.97% and 64.01%, respectively. This represents the best performance, demonstrating both robustness and generalization capabilities. Consequently, it offers a reliable automated detection method for producing high-quality welded products.

Wheat Flour Discrimination Using Two-Dimensional Correlation Spectroscopy and Deep Learning.

The continuous evolution of deep learning has garnered significant attention in spectroscopy. This study focuses on identifying wheat flour, presenting a more efficient and accurate method by combining two-dimensional correlation spectroscopy (2D-COS) and deep learning techniques. A data set of 316 near-infrared (NIR) spectral samples of four types of wheat flour was collected. By applying three disparate 2D-COS techniques, i.e., synchronous, asynchronous, and integrated, we crafted 948 2D-COS images. These images, obtained by transforming the original one-dimensional spectra into 2D representations, offer richer information for deep learning analysis. The study introduced an 18-layer residual network incorporating a convolutional attention mechanism, specifically tailored for the 2D-COS analysis of wheat flour, aimed at enhancing the model's discriminative capabilities by refining the residual neural network's structure. Achieving an unprecedented recognition accuracy of 100% through methodical optimization and rigorous training on the synchronous 2D-COS data set of wheat flour, the proposed model is a testament to the efficacy of deep learning in spectroscopic analysis. To further exhibit the confluence of 2D-COS with deep learning, t-distributed stochastic neighbor embedding was employed to visualize the distinctive 2D-COS features within the deep learning architecture. Additionally, the model's performance was juxtaposed with prevailing NIR spectral recognition methods, including random forest, gradient boosting decision tree, and artificial neural network. This comparison cemented the proposed approach's superiority in wheat flour categorization. The findings of this study not only introduce a novel and efficient solution for wheat flour quality analysis but also underscore the significant potential of deep learning techniques in spectroscopy applications.

Research on high-precision recognition model for multi-scene asphalt pavement distresses based on deep learning

Accurate detection of asphalt pavement distress is crucial for road maintenance and traffic safety. However, traditional convolutional neural networks usually struggle with this task due to the varied distress patterns and complex background in the images. To enhance the accuracy of asphalt pavement distress identification across various scenarios, this paper introduces an improved model named SMG-YOLOv8, based on the YOLOv8s framework. This model integrates the space-to-depth module and the multi-scale convolutional attention mechanism, while optimizing the backbone’s C2f structure with a more efficient G-GhostC2f structure. Experimental results demonstrate that SMG-YOLOv8 outperforms the YOLOv8s baseline model, achieving Pmacro and mAP50 scores of 81.1% and 79.4% respectively, marking an increase of 8.2% and 12.5% over the baseline. Furthermore, SMG-YOLOv8 exhibits clear advantages in identifying various types of pavement distresses, including longitudinal cracks, transverse cracks, mesh cracks, and potholes, when compared to YOLOv5n, YOLOv5s, YOLOv6s, YOLOv8n, and SSD models. This enhancement optimizes the network structure, reducing the number of parameters while maintaining excellent detection performance. In real-world scenarios, the SMG-YOLOv8 model, when applied to image data collected from projects, achieves a Pmacro of 80.5% and an Rmacro of 86.2%. This result demonstrates its excellent generalization capability and practicality. The model provides significant technical support for the intelligent detection of pavement distress.

An Improved Multi-Scale Feature Extraction Network for Rice Disease and Pest Recognition

In the process of rice production, rice pests are one of the main factors that cause rice yield reduction. To implement prevention and control measures, it is necessary to accurately identify the types of rice pests and diseases. However, the application of image recognition technologies focused on the agricultural field, especially in the field of rice disease and pest identification, is relatively limited. Existing research on rice diseases and pests has problems such as single data types, low data volume, and low recognition accuracy. Therefore, we constructed the rice pest and disease dataset (RPDD), which was expanded through data enhancement methods. Then, based on the ResNet structure and the convolutional attention mechanism module, we proposed a Lightweight Multi-scale Feature Extraction Network (LMN) to extract multi-scale features at a finer granularity. The proposed LMN model achieved an average classification accuracy of 95.38% and an F1-Score of 94.5% on the RPDD. The parameter size of the model is 1.4 M, and the FLOPs is 1.65 G. The results suggest that the LMN model performs rice disease and pest classification tasks more effectively than the baseline ResNet model by significantly reducing the model size and improving accuracy.

Semantic Segmentation of Satellite Images for Landslide Detection Using Foreground-Aware and Multi-Scale Convolutional Attention Mechanism.

Advancements in satellite and aerial imagery technology have made it easier to obtain high-resolution remote sensing images, leading to widespread research and applications in various fields. Remote sensing image semantic segmentation is a crucial task that provides semantic and localization information for target objects. In addition to the large-scale variation issues common in most semantic segmentation datasets, aerial images present unique challenges, including high background complexity and imbalanced foreground-background ratios. However, general semantic segmentation methods primarily address scale variations in natural scenes and often neglect the specific challenges in remote sensing images, such as inadequate foreground modeling. In this paper, we present a foreground-aware remote sensing semantic segmentation model. The model introduces a multi-scale convolutional attention mechanism and utilizes a feature pyramid network architecture to extract multi-scale features, addressing the multi-scale problem. Additionally, we introduce a Foreground-Scene Relation Module to mitigate false alarms. The model enhances the foreground features by modeling the relationship between the foreground and the scene. In the loss function, a Soft Focal Loss is employed to focus on foreground samples during training, alleviating the foreground-background imbalance issue. Experimental results indicate that our proposed method outperforms current state-of-the-art general semantic segmentation methods and transformer-based methods on the LS dataset benchmark.

MG-Net: A fetal brain tissue segmentation method based on multiscale feature fusion and graph convolution attention mechanisms

Background and Objective:Fetal brain tissue segmentation provides foundational support for comprehensively understanding the neurodevelopment of normal and congenital disease-affected fetuses. Manual labeling is very time-consuming, and automated segmentation methods can greatly improve the efficiency of doctors. At the same time, fetal brain tissue undergoes various changes throughout the pregnancy, leading to a continuous change in tissue contrast, which greatly increases the difficulty of training segmentation methods. This study aims to develop an automated segmentation model that can efficiently and accurately segment fetal brain tissue, improving the workflow for medical professionals. Methods:We propose a novel deep learning-based segmentation model that incorporates three innovative components: Firstly, a new Dual Dilated Attention Block (DDAB) is proposed in the encoder part to enhance the feature extraction of local spatial and structural contextual information. Secondly, a Multi-scale Deformable Transformer (MSDT) is integrated into the bottleneck to improve the feature extraction of global information on local spatial and structural contextual information. Thirdly, we use a novel block based on Graph Convolution Attention (GCAB) in the decoder, which effectively enhances the features at the decoder.The code is available at https://github.com/unicoco7/MG-Net/. Results:We trained and tested on the FeTA 2021 and FeTA 2022 datasets, and evaluated using seven popular metrics, including Dice, IoU, MAE, BoundaryF, PRE, SEN, and SPE. Compared to the current state-of-the-art 3D segmentation models such as nnFormer, SwinUNETR, and 3DUX-net, our proposed method has surpassed all of them in metrics like Dice, IoU, and MAE. Specifically, on the FeTA 2021 dataset, our model achieved a Dice of 0.8666, an IoU of 0.7646, and an MAE of 0.0027; on the FeTA 2022 dataset, it achieved a Dice of 0.8552, an IoU of 0.7470, and an MAE of 0.0005. Conclusion:In this paper, we propose a model for three-dimensional fetal brain tissue segmentation based on multi-scale feature fusion and graph convolution attention mechanism, and conduct experimental evaluation on the FeTA 2021 and FeTA 2022 datasets. Understanding the boundaries of fetal brain tissue is crucial for doctors’ diagnosis, so the proposed model is expected to improve the speed and accuracy of doctors’ diagnoses.

Boundary-aware convolutional attention network for liver segmentation in ultrasound images

Liver ultrasound is widely used in clinical practice due to its advantages of non-invasiveness, non-radiation, and real-time imaging. Accurate segmentation of the liver region in ultrasound images is essential for accelerating the auxiliary diagnosis of liver-related diseases. This paper proposes BACANet, a deep learning algorithm designed for real-time liver ultrasound segmentation. Our approach utilizes a lightweight network backbone for liver feature extraction and incorporates a convolutional attention mechanism to enhance the network's ability to capture global contextual information. To improve early localization of liver boundaries, we developed a selective large kernel convolution module for boundary feature extraction and introduced explicit liver boundary supervision. Additionally, we designed an enhanced attention gate to efficiently convey liver body and boundary features to the decoder to enhance the feature representation capability. Experimental results across multiple datasets demonstrate that BACANet effectively completes the task of liver ultrasound segmentation, achieving a balance between inference speed and segmentation accuracy. On a public dataset, BACANet achieved a DSC of 0.921 and an IOU of 0.854. On a private test dataset, BACANet achieved a DSC of 0.950 and an IOU of 0.907, with an inference time of approximately 0.32 s per image on a CPU processor.

US-Net: U-shaped network with Convolutional Attention Mechanism for ultrasound medical images

Ultrasound imaging, characterized by low contrast, high noise, and interference from surrounding tissues, poses significant challenges in lesion segmentation. To tackle these issues, we introduce an enhanced U-shaped network that incorporates several novel features for precise, automated segmentation. Firstly, our model utilizes a convolution-based self-attention mechanism to establish long-range dependencies in feature maps, crucial for small dataset applications, accompanied by a soft thresholding method for noise reduction. Secondly, we employ multi-sized convolutional kernels to enrich feature processing, coupled with curvature calculations to accentuate edge details via a soft-attention approach. Thirdly, an advanced skip connection strategy is implemented in the UNet architecture, integrating information entropy to assess and utilize texture-rich channels, thereby improving semantic detail in the encoder before merging with decoder outputs. We validated our approach using a newly curated dataset, VPUSI (Vascular Plaques Ultrasound Images), alongside the established datasets, BUSI, TN3K and DDTI. Comparative experiments on these datasets show that our model outperforms existing state-of-the-art techniques in segmentation accuracy.

Effective lane detection on complex roads with convolutional attention mechanism in autonomous vehicles

Autonomous Vehicles (AV’s) have achieved more popularity in vehicular technology in recent years. For the development of secure and safe driving, these AV’s help to reduce the uncertainties such as crashes, heavy traffic, pedestrian behaviours, random objects, lane detection, different types of roads and their surrounding environments. In AV’s, Lane Detection is one of the most important aspects which helps in lane holding guidance and lane departure warning. From Literature, it is observed that existing deep learning models perform better on well maintained roads and in favourable weather conditions. However, performance in extreme weather conditions and curvy roads need focus. The proposed work focuses on presenting an accurate lane detection approach on poor roads, particularly those with curves, broken lanes, or no lane markings and extreme weather conditions. Lane Detection with Convolutional Attention Mechanism (LD-CAM) model is proposed to achieve this outcome. The proposed method comprises an encoder, an enhanced convolution block attention module (E-CBAM), and a decoder. The encoder unit extracts the input image features, while the E-CBAM focuses on quality of feature maps in input images extracted from the encoder, and the decoder provides output without loss of any information in the original image. The work is carried out using the distinct data from three datasets called Tusimple for different weather condition images, Curve Lanes for different curve lanes images and Cracks and Potholes for damaged road images. The proposed model trained using these datasets showcased an improved performance attaining an Accuracy of 97.90%, Precision of 98.92%, F1-Score of 97.90%, IoU of 98.50% and Dice Co-efficient as 98.80% on both structured and defective roads in extreme weather conditions.

An efficient steel defect detection model based on multi-scale information extraction

PurposeIn the field of steel defect detection, the existing detection algorithms struggle to achieve a satisfactory balance between detection accuracy, computational cost and inference speed due to the interference from complex background information, the variety of defect types and significant variations in defect morphology. To solve this problem, this paper aims to propose an efficient detector based on multi-scale information extraction (MSI-YOLO), which uses YOLOv8s as the baseline model.Design/methodology/approachFirst, the authors introduce an efficient multi-scale convolution with different-sized convolution kernels, which enables the feature extraction network to accommodate significant variations in defect morphology. Furthermore, the authors introduce the channel prior convolutional attention mechanism, which allows the network to focus on defect areas and ignore complex background interference. Considering the lightweight design and accuracy improvement, the authors introduce a more lightweight feature fusion network (Slim-neck) to improve the fusion effect of feature maps.FindingsMSI-YOLO achieves 79.9% mean average precision on the public data set Northeastern University (NEU)-DET, with a model size of only 19.0 MB and an frames per second of 62.5. Compared with other state-of-the-art detectors, MSI-YOLO greatly improves the recognition accuracy and has significant advantages in computational cost and inference speed. Additionally, the strong generalization ability of MSI-YOLO is verified on the collected industrial site steel data set.Originality/valueThis paper proposes an efficient steel defect detector with high accuracy, low computational cost, excellent detection speed and strong generalization ability, which is more valuable for practical applications in resource-limited industrial production.

Real-time topology optimization based on multi-scale convolutional attention mechanism

This article proposes a deep learning model based on a multi-scale convolutional attention mechanism network SegNext for improving Topology Optimization (TO-Next), aimed at addressing the computational challenges associated with finite element iterations in traditional density-based methods. TO-Next is trained on diverse topology structures with varying loads, constraints and enriched physical information after initial iterations to enhance its generalization. Following training on three distinct decoder architectures, the optimal encoder–decoder network structure is determined. This study also investigates the algorithm's generalization capability under both single and multiple load constraints. The results indicate the superior performance and efficiency of the proposed method compared with the Solid Isotropic Material with Penalization (SIMP) method, enabling real-time generation of near-optimal topology structures within a short timeframe.

The damage localization method in the composite panel by multifrequency guided wave signal using parallel network

Fiber-reinforced composite materials find extensive application in the realm of aeronautics; nonetheless, during operational phases, structural integrity may be compromised due to external impacts or extreme environmental conditions. Consequently, the imperative is the necessity of vigilantly monitoring the structural state. The structural health monitoring (SHM) technology based on ultrasonic guided waves can achieve long-distance and wide-ranging damage detection utilizing a minimal number of cost-effective piezoelectric transducers. However, traditional damage identification methods necessitate the analysis of the multi-modal characteristics of guided waves and the deployment of dense sensor arrays. In this paper, we propose a network model that leverages multi-frequency guided-wave signals to achieve spatiotemporal parallelism, incorporating residual blocks and convolutional attention mechanisms (GCRA). This model accurately estimates damage locations with an $R^2$ value of 0.9147. The average error for the six randomly selected damage locations is 0.4275 cm. The results demonstrate that the GCRA facilitates the identification of damaged positions with a high degree of confidence.

RICE: A Dataset and Baseline for Cloud Removal in Remote Sensing Images

Removing clouds is an essential preprocessing step in analyzing remote sensing images, as cloud-based overlays commonly occur in optical remote sensing images and can significantly limit the usability of the acquired data. Deep learning has exhibited remarkable progress in remote sensing, encompassing scene classification and change detection tasks. Nevertheless, the appli-cation of deep learning techniques to cloud removal in remote sensing images is currently con-strained by the limited availability of training datasets explicitly tailored for neural networks. This paper presents the Remote sensing Image Cloud rEmoving dataset (RICE) to address this challenge and proposes baseline models incorporating a convolutional attention mechanism, which has demonstrated superior performance in identifying and restoring cloud-affected regions, with quantitative results indicating a 3.08% improvement in accuracy over traditional methods. This mechanism empowers the network to comprehend better the spatial structure, local details, and inter-channel correlations within remote sensing images, thus effectively addressing the diverse distributions of clouds. Moreover, by integrating this attention mechanism, our models achieve a crucial comparison advantage, outperforming existing state-of-the-art techniques in terms of both visual quality and quantitative metrics. We propose adopting the Learned Per-ceptual Image Patch Similarity metric, which emphasizes perceptual similarity, to evaluate the quality of cloud-free images generated by the models. Our work not only contributes to advancing cloud removal techniques in remote sensing but also provides a comprehensive evaluation framework for assessing the fidelity of the generated images.

NnSegNeXt: A 3D Convolutional Network for Brain Tissue Segmentation Based on Quality Evaluation.

Accurate and automated segmentation of brain tissue images can significantly streamline clinical diagnosis and analysis. Manual delineation needs improvement due to its laborious and repetitive nature, while automated techniques encounter challenges stemming from disparities in magnetic resonance imaging (MRI) acquisition equipment and accurate labeling. Existing software packages, such as FSL and FreeSurfer, do not fully replace ground truth segmentation, highlighting the need for an efficient segmentation tool. To better capture the essence of cerebral tissue, we introduce nnSegNeXt, an innovative segmentation architecture built upon the foundations of quality assessment. This pioneering framework effectively addresses the challenges posed by missing and inaccurate annotations. To enhance the model's discriminative capacity, we integrate a 3D convolutional attention mechanism instead of conventional convolutional blocks, enabling simultaneous encoding of contextual information through the incorporation of multiscale convolutional features. Our methodology was evaluated on four multi-site T1-weighted MRI datasets from diverse sources, magnetic field strengths, scanning parameters, temporal instances, and neuropsychiatric conditions. Empirical evaluations on the HCP, SALD, and IXI datasets reveal that nnSegNeXt surpasses the esteemed nnUNet, achieving Dice coefficients of 0.992, 0.987, and 0.989, respectively, and demonstrating superior generalizability across four distinct projects with Dice coefficients ranging from 0.967 to 0.983. Additionally, extensive ablation studies have been implemented to corroborate the effectiveness of the proposed model. These findings represent a notable advancement in brain tissue analysis, suggesting that nnSegNeXt holds the promise to significantly refine clinical workflows.

Identification of Radio Frequency Interference Using Multi-scale TransUNet

Radio observation is a method for conducting astronomical observations using radio waves. A common challenge in radio observations is Radio Frequency Interference (RFI), which refers to the unintentional or intentional interference of radio signals from other wireless sources within the radio frequency band. Such interference contaminates the astronomical signals received by radio telescopes, significantly affecting time–frequency domain astronomical observations and research. Consequently, identifying RFI is crucial. In this paper, we employ a deep learning approach to detect RFI present in observation data and propose an improved network structure based on TransUNet. This network leverages the principles of a multi-scale convolutional attention mechanism. It introduces an auxiliary branch to extract high-dimensional image information and an enhanced coordinate attention mechanism for feature map extraction, enabling more comprehensive and accurate identification of RFI in time–frequency images. We introduce a novel architecture named the Multi-Scale TransUNet Network, abbreviated as MS-TransUNet. We utilized observation data from the 40 m radio telescope at the Yunnan Observatory as a data set for training, validating, and testing the network. Compared with previous deep learning networks (U-Net, RFI-Net, R-Net, DSC, EMSCA-UNet), the recall rate and f2 score have been significantly improved. Specifically, the recall rate is improved by at least 2.99%, and the f2 score is improved by at least 2.46%. Experiments demonstrate that this network is exceptional in identifying RFI more comprehensively while ensuring high precision.

Infrared image fault diagnosis based on dual-stream attention convolution network

The development of infrared image detection technology has improved the real-time performance and safety of fault diagnosis. To address the problem of insufficient features caused by the low resolution of infrared images, a fault diagnosis model for induction motor infrared images based on Dual-Stream Attention Convolutional (DSAC) is proposed. Firstly, this model employs a dual-stream convolutional neural network to extract spatial features in the X and Y directions of the infrared images separately. This dual-stream structure allows the network to simultaneously learn and utilize information from both X and Y directions, enabling a more comprehensive capture of temperature distribution and variation trends in the infrared images. Then, a convolutional attention mechanism is introduced to assign weights to the obtained features. The attention maps generated by the convolutional attention layer enhance key features while suppressing unimportant information, thereby enhancing the focusing performance of the DSAC model on critical features. Finally, the weighted dual-stream features are fused to achieve the goal of fault diagnosis for induction motor infrared images. The DSAC model is validated using the induction motor infrared image dataset from Babol Noshirvani University of Technology, demonstrating excellent diagnostic accuracy and speed on small sample datasets, providing a feasible solution for fault diagnosis of induction motor infrared images.