Residual Convolutional Neural Network Research Articles

Innovative design of metamaterial perfect absorbers via residual fully connected neural network modeling

In recent years, significant progress has been made in the research of metamaterial perfect absorbers for on-demand design based on deep learning methods, and how to improve the generalization ability and prediction accuracy of network models in inverse design has been a hot topic of recent research. In this paper, we propose a fully connected neural network (RFC-NN) based on the residual principle applied to the inverse design of perfect absorbers. We design the RFC-NN model by applying the residual network structure commonly used on convolutional neural networks to a fully connected neural network, and the regression coefficients of the predicted structural parameters and the corresponding absorption spectra in the inverse design are 0.988 and 0.923, respectively, which have better prediction performance and generalization ability than the ordinary fully connected neural network (FC-NN), tandem neural network (TNN), and one-dimensional residual convolutional neural network (1D-Resnet). Then, we successfully designed the metamaterial perfect absorber with an absorption bandwidth (the spectral range where the absorptivity is greater than 90%) of 1935 nm using the trained RFC-NN model, which corresponds to an average absorptivity of 93.70%. Meanwhile, we use the analysis method of the Pearson correlation coefficient to improve the accuracy of the RFC-NN model after inverse design. Our proposed design method proves to be very effective and can also be applied to the design of other types of functional nanophotonic devices.

CLASSIFICATION OF PARKINSON'S DISEASE IN BRAIN MRI IMAGES USING DEEP RESIDUAL CONVOLUTIONAL NEURAL NETWORK

In our aging culture, neurodegenerative disorders like Parkinson's disease (PD) are among the most serious health issues. It is a neurological condition that has social and economic effects on individuals. It happens because the brain's dopamine-producing cells are unable to produce enough of the chemical to support the body's motor functions. The main symptoms of this illness are eyesight, excretion activity, speech, and mobility issues, followed by depression, anxiety, sleep issues, and panic attacks. The main aim of this research is to develop a workable clinical decision-making framework that aids the physician in diagnosing patients with PD influence. In this research, we proposed a technique to classify Parkinson’s disease by MRI brain images. Initially, normalize the input data using the min-max normalization method and then remove noise from input images using a median filter. Then utilizing the Binary Dragonfly Algorithm to select the features. Furthermore, to segment the diseased part from MRI brain images using the technique Dense-UNet. Then, classify the disease as if it’s Parkinson’s disease or health control using the Deep Residual Convolutional Neural Network (DRCNN) technique along with Enhanced Whale Optimization Algorithm (EWOA) to get better classification accuracy. Here, we use the public Parkinson’s Progression Marker Initiative (PPMI) dataset for Parkinson’s MRI images. The accuracy, sensitivity, specificity, and precision metrics will be utilized with manually gathered data to assess the efficacy of the proposed methodology.

Learning a Deep Attention Dilated Residual Convolutional Neural Network for Landslide Susceptibility Mapping in Hanzhong City, Shaanxi Province, China

The analysis and evaluation of landslide susceptibility are of great significance in preventing and managing geological hazards. Aiming at the problems of insufficient information caused by the limited number of landslide datasets, complex information of landslide evaluation factors, and low prediction accuracy of landslide susceptibility, a landslide susceptibility evaluation method based on the deep attention dilated residual convolutional neural network (DADRCNN) is proposed. First, the dilated convolution unit (DCU) is used to increase the network receptive field, aggregate multi-scale information, and enhance the model ability to capture the characteristics of landslide evaluation factors. Second, the deep residual module (DRM) is used to solve the issue of gradient disappearance and better extract data features by overlaying the residual function mapping layer and increasing the network depth. Finally, the channel attention residual module (CARM) is introduced to learn the varying importance of different landslide evaluation factors, and assign different weights to improve the susceptibility prediction accuracy. The experimental results show that the DADRCNN method can extract features around the sample points, expand the receptive field, and deeply mine the information. It mitigates the lack of sample information in training, focuses on important feature information, and significantly improves the prediction accuracy.

Prediction of fat content in salmon fillets based on hyperspectral imaging and residual attention convolution neural network

The fat content of salmon is an essential indicator to evaluate its quality, and is related to its commercial value. An attention residual convolutional neural network (CNN), namely RACNN, which added channel attention and residual modules to the traditional CNN structure, was proposed to combine near-infrared hyperspectral imaging (NIR-HSI) (900-1700 nm) to predict fat contents in salmon fillets. Partial least squares regression (PLSR), least squares support vector machine (LSSVM) and CNN models were compared with the proposed RACNN models. Various spectrum preprocessing methods were used at the time. And Three feature extraction algorithms (successive projections algorithm (SPA), uninformative variable elimination (UVE), and competitive adaptive reweighted sampling (CARS)) were used to extract the effective wavelengths from the preprocessed NIR spectra. The experimental results showed that, the optimal PLSR and LSSVM models obtained coefficient of determination of prediction (Rp2) values of 0.8781 and 0.8952, respectively. The RACNN model based on full spectrum gained the highest prediction accuracies with Rp2 = 0.9033, root mean square error of prediction (RMSEP) = 1.5143, and residual predictive deviation (RPD) = 3.2707. Therefore, the proposed RACNN model enables the rapid, non-destructive and accurate prediction of the fat content in salmon fillets without additional processing on raw spectra.

Multi-modal deep learning for automated assembly of periapical radiographs

ObjectivesPeriapical radiographs are oftentimes taken in series to display all teeth present in the oral cavity. Our aim was to automatically assemble such a series of periapical radiographs into an anatomically correct status using a multi-modal deep learning model. Methods4,707 periapical images from 387 patients (on average, 12 images per patient) were used. Radiographs were labeled according to their field of view and the dataset split into a training, validation, and test set, stratified by patient. In addition to the radiograph the timestamp of image generation was extracted and abstracted as follows: A matrix, containing the normalized timestamps of all images of a patient was constructed, representing the order in which images were taken, providing temporal context information to the deep learning model. Using the image data together with the time sequence data a multi-modal deep learning model consisting of two residual convolutional neural networks (ResNet-152 for image data, ResNet-50 for time data) was trained. Additionally, two uni-modal models were trained on image data and time data, respectively. A custom scoring technique was used to measure model performance. ResultsMulti-modal deep learning outperformed both uni-modal image-based learning (p<0.001) and time-based learning (p<0.05). The multi-modal deep learning model predicted tooth labels with an F1-score, sensitivity and precision of 0.79, respectively, and an accuracy of 0.99. 37 out of 77 patient datasets were fully correctly assembled by multi-modal learning; in the remaining ones, usually only one image was incorrectly labeled. ConclusionsMulti-modal modeling allowed automated assembly of periapical radiographs and outperformed both uni-modal models. Dental machine learning models can benefit from additional data modalities. Clinical significanceLike humans, deep learning models may profit from multiple data sources for decision-making. We demonstrate how multi-modal learning can assist assembling periapical radiographs into an anatomically correct status. Multi-modal learning should be considered for more complex tasks, as clinically a wealth of data is usually available and could be leveraged.

AB-MTEDeep Classifier Trained with AAGAN for the Identification and Classification of Alopecia Areata

Artificial Intelligence (AI) is widely used in dermatology to analyze trichoscopy imaging and assess Alopecia Areata (AA) and scalp hair problems. From this viewpoint, the Attention-based Balanced Multi-Tasking Ensembling Deep (AB-MTEDeep) network was developed, which combined the Faster Residual Convolutional Neural Network (FRCNN) and Long Short-Term Memory (LSTM) network with cross residual learning to classify scalp images into different AA classes. This article presents a new data augmentation model called AA-Generative Adversarial Network (AA-GAN) to produce a huge number of images from a set of input images. The structure of AA-GAN and its loss functions are comparable to those of standard GAN, which encompasses a generator and a discriminator network. To generate high-quality AA structure-based images, the generator was trained to extract the 2D orientation and confidence maps along with the bust depth map from real hair and scalp images. The discriminator was also used to separate real from generated images, which were provided as feedback to the generator to create synthetic images that are extremely close to the real input images. The created images were used to train the AB-MTEDeep model for AA classification. Finally, the experimental results exhibited that the AA-GAN-AB-MTEDeep achieved 96.94% accuracy.

Nonspeech7k dataset: Classification and analysis of human non‐speech sound

AbstractHuman non‐speech sounds occur during expressions in a real‐life environment. Realising a person's incapability to prompt confident expressions by non‐speech sounds may assist in identifying premature disorder in medical applications. A novel dataset named Nonspeech7k is introduced that contains a diverse set of human non‐speech sounds, such as the sounds of breathing, coughing, crying, laughing, screaming, sneezing, and yawning. The authors then conduct a variety of classification experiments with end‐to‐end deep convolutional neural networks (CNN) to show the performance of the dataset. First, a set of typical deep classifiers are used to verify the reliability and validity of Nonspeech7k. Involved CNN models include 1D‐2D deep CNN EnvNet, deep stack CNN M11, deep stack CNN M18, intense residual block CNN ResNet34, modified M11 named M12, and the authors’ baseline model. Among these, M12 achieves the highest accuracy of 79%. Second, to verify the heterogeneity of Nonspeech7k with respect to two typical datasets, FSD50K and VocalSound, the authors design a series of experiments to analyse the classification performance of deep neural network classifier M12 by using FSD50K, FSD50K + Nonspeech7k, VocalSound, VocalSound + Nonspeech7k as training data, respectively. Experimental results show that the classifier trained with existing datasets mixed with Nonspeech7k achieves the highest accuracy improvement of 15.7% compared to that without Nonspeech7k mixed. Nonspeech7k is 100% annotated, completely checked, and free of noise. It is available at https://doi.org/10.5281/zenodo.6967442.

Joint demosaicking and denoising benefits from a two-stage training strategy

Image demosaicking and denoising are the first two key steps of the color image production pipeline. The classical processing sequence has for a long time consisted of applying denoising first, and then demosaicking. Applying the operations in this order leads to oversmoothing and checkerboard effects. Yet, it was difficult to change this order, because once the image is demosaicked, the statistical properties of the noise are dramatically changed and hard to handle by traditional denoising models. In this paper, we address this problem by a hybrid machine learning method. We invert the traditional color filter array (CFA) processing pipeline by first demosaicking and then denoising. Our demosaicking algorithm, trained on noiseless images, combines a traditional method and a residual convolutional neural network (CNN). This first stage retains all known information, which is the key point to obtain faithful final results. The noisy demosaicked image is then passed through a second CNN restoring a noiseless full-color image. This pipeline order completely avoids checkerboard effects and restores fine image detail. Although CNNs can be trained to solve jointly demosaicking–denoising end-to-end, we find that this two-stage training performs better and is less prone to failure. It is shown experimentally to improve on the state of the art, both quantitatively and in terms of visual quality.

Spectroscopic analysis improvement using convolutional neural networks

Removing noisy components of signals coming from edge tokamak plasmas, astrochemical organic matter or astronomical objects is one of the promising path to improve the underlying elements identification. Methods such as the penalized semi-supervised non negative matrix factorization (PSNMF) used to extract such prominent elements perform well on complex signals. However, it is results’ confidence decreases as the noise increases. In this context, we have tried to address this limitation by removing part of the undesired noise in atomic spectra using artificial intelligence (AI) method based on convolutional neural networks (CNNs). More specifically, we have tested different architectures of CNN classically used in denoising task: residual CNNs and auto-encoders, to benchmark their respective denoising capacity. The dataset used is made of high resolution atom and ion spectra extracted from the NIST Atomic Spectra Database. In the case of ions, we have considered several degrees of ionization. The synthetic added noise is generated from a typical experimental noise profile randomly modified for each signal. Performance of each AI-based noise remover is measured by analyzing the increase of the element identification precision obtained by the PSNMF. Auto-encoders and residual CNN are both suitable to remove the undesired noise and increase the PSNMF efficiency even for noise-prevailing signals. In this particular case, the auto-encoder architecture seems to be globally more accurate and should be selected when considering noisy multi-element atomic spectra.

MARC-Net: Terrain Classification in Parallel Network Architectures Containing Multiple Attention Mechanisms and Multi-Scale Residual Cascades

To address the problem that traditional deep learning algorithms cannot fully utilize the correlation properties between spectral sequence information and the feature differences between different spectra, this paper proposes a parallel network architecture land-use classification based on a combined multi-head attention mechanism and multiscale residual cascade called MARC-Net. This parallel framework is firstly implemented by deeply mining the features generated by grouped spectral embedding for information among spectral features by adding a multi-head attention mechanism, which allows semantic features to have expressions from more subspaces while fully considering all spatial location interrelationships. Secondly, a multiscale residual cascade CNN (convolutional neural network) is designed to fully utilize the fused feature information at different scales, thereby improving the network’s ability to represent different levels of information. Lastly, the features obtained by the multi-head attention mechanism are fused with those obtained by the CNN, and the merged resultant features are downgraded through the fully connected layer to obtain the classification results and achieve pixel-level multispectral image classification. The findings show that the algorithm proposed in this paper has an aggregate precision of 97.22%, compared to that of the Vision Transformer (ViT) with 95.08%; its performance on the Sentinel-2 dataset shows a huge improvement. Moreover, this article mainly focuses on the change rate of forest land in the study area. The Forest land area was 125.1143 km2 in 2017, 105.6089 km2 in 2019, and 76.3699 km2 in 2021, with an increase of 15.59%, an decrease of 0.97%, and increase of 14.76% in 2017–2019, 2019–2021 and 2017–2021, respectively.

Semantic segmentation of retinal exudates using a residual encoder-decoder architecture in diabetic retinopathy.

Exudates are a common sign of diabetic retinopathy, which is a disease that affects the blood vessels in the retina. Early detection of exudates is critical to avoiding vision problems through continuous screening and treatment. In traditional clinical practice, the involved lesions are manually detected using photographs of the fundus. However, this task is cumbersome and time-consuming and requires intense effort due to the small size of the lesion and the low contrast of the images. Thus, computer-assisted diagnosis of retinal disease based on the detection of red lesions has been actively explored recently. In this paper, we present a comparison of deep convolutional neural network (CNN) architectures and propose a residual CNN with residual skip connections to reduce the parameter for the semantic segmentation of exudates in retinal images. A suitable image augmentation technique is used to improve the performance of network architecture. The proposed network can robustly segment exudates with high accuracy, which makes it suitable for diabetic retinopathy screening. A comparative performance analysis of three benchmark databases: E-ophtha, DIARETDB1, and Hamilton Ophthalmology Institute's Macular Edema, is presented. The proposed method achieves a precision of 0.95, 0.92, 0.97, accuracy of 0.98, 0.98, 0.98, sensitivity of 0.97, 0.95, 0.95, specificity of 0.99, 0.99, 0.99, and area under the curve of 0.97, 0.94, and 0.96, respectively. RESEARCH HIGHLIGHTS: The research focuses on the detection and segmentation of exudates in diabetic retinopathy, a disease affecting the retina. Early detection of exudates is important to avoid vision problems and requires continuous screening and treatment. Currently, manual detection is time-consuming and requires intense effort. The authors compare qualitative results of the state-of-the-art convolutional neural network (CNN) architectures and propose a computer-assisted diagnosis approach based on deep learning, using a residual CNN with residual skip connections to reduce parameters. The proposed method is evaluated on three benchmark databases and demonstrates high accuracy and suitability for diabetic retinopathy screening.

Knowledge guided multi-filter residual convolutional neural network for ICD coding from clinical text

A common challenge encountered when using Deep Neural Network models for automatic ICD coding is their potential inability to effectively handle unseen clinical texts, especially when these models are only trained on a limited number of examples. This is because these models rely solely on the patterns and relationships present in the training data, and may not be able to effectively incorporate additional knowledge about the relationships between medical entities. To address this issue, we introduce KG-MultiResCNN—KnowledgeGuidedMulti-filterResidualConvolutionalNeuralNetwork model, which combines training examples with external knowledge from the Wikidata Knowledge Graph (KG) in order to better capture the relationships between medical entities. The KG is a structured database that contains a wealth of information about various entities, including medical concepts and their relationships with one another. By incorporating this external knowledge into our model, we are able to improve its ability to predict ICD codes for new clinical texts. In our experiments with the MIMIC-III dataset, we found that the KG-MultiResCNN model significantly outperformed the baseline approaches. This demonstrates the effectiveness of using external knowledge, in addition to training examples, to improve the performance of deep learning models for automatic ICD coding.

A trial deep learning-based model for four-class histologic classification of colonic tumor from narrow band imaging

Narrow band imaging (NBI) has been extensively utilized as a diagnostic tool for colorectal neoplastic lesions. This study aimed to develop a trial deep learning (DL) based four-class classification model for low-grade dysplasia (LGD); high-grade dysplasia or mucosal carcinoma (HGD); superficially invasive submucosal carcinoma (SMs) and deeply invasive submucosal carcinomas (SMd) and evaluate its potential as a diagnostic tool. We collected a total of 1,390 NBI images as the dataset, including 53 LGD, 120 HGD, 20 SMs and 17 SMd. A total of 598,801 patches were trimmed from the lesion and background. A patch-based classification model was built by employing a residual convolutional neural network (CNN) and validated by three-fold cross-validation. The patch-based validation accuracy was 0.876, 0.957, 0.907 and 0.929 in LGD, HGD, SMs and SMd, respectively. The image-level classification algorithm was derived from the patch-based mapping across the entire image domain, attaining accuracies of 0.983, 0.990, 0.964, and 0.992 in LGD, HGD, SMs, and SMd, respectively. Our CNN-based model demonstrated high performance for categorizing the histological grade of dysplasia as well as the depth of invasion in routine colonoscopy, suggesting a potential diagnostic tool with minimal human inputs.

General-purpose mid-infrared micro-spectrometer based on hierarchical residual CNN and data augmentation.

Taking advantage of broad response range and snap-shot operation mode, reconstructive spectrometers based on integrated frequency-modulation microstructure and computational techniques attract lots of attention. The key problems in reconstruction are sparse samplings related with the limited detectors and generalization ability due to data-driving principle. Here, we demonstrate abstractly a mid-infrared micro-spectrometer covering 2.5-5 μm, which utilizes a grating-integrated lead selenide detector array for sampling and a hierarchal residual convolutional neural network (HRCNN) for reconstructions. Leveraging data augmentation and the powerful feature extraction ability of HRCNN, a spectral resolution of 15 nm is realized. Over one hundred chemicals, including untrained chemicals species tested with an average reconstruction error of ∼1E-4, exhibit the excellent reliability of the micro-spectrometer. The demonstration of the micro-spectrometer promotes the development of the reconstructed strategy.

Improved burned area mapping using monotemporal Landsat-9 imagery and convolutional shift-transformer

Satellite imagery, specifically Landsat, have been widely used for mapping and monitoring wildfire burned areas. The new Landsat-9 satellite – with higher radiometric resolution compared to its predecessors, and improved temporal resolution when combined with Landsat-8 (∼8 days) – enables a wide range of applications, particularly burned area mapping (BAM). We propose a novel deep learning BAM model that leverages the strengths of the convolutional layers for deep feature generation from Landsat-9 imagery and shift-transformer block for burned area classification. The performance of the model is evaluated in five large fire case studies across the globe. BAM results are also compared with two state-of-the-art models, namely residual convolutional neural network and vision transformer. The proposed convolutional shift-transformer (CST) outperforms other models with an F1-score of greater than 96% across the case studies. Furthermore, CST only requires a single post-fire image that reduces computational costs compared to traditional models that use bi-temporal images.

Automated classification of pediatric age from 12-lead ECG using deep residual convolutional neural networks

With increased interest in screening of young people for potential causes of sudden death, accurate automated detection of ventricular pre-excitation (VPE) or Wolff–Parkinson–White syndrome (WPW) in the pediatric resting ECG is important. Several recent studies have shown interobserver variability when reading screening ECGs and thus an accurate automated reading for this potential cause of sudden death is critical. We designed and tested an automated algorithm to detect pediatric VPE optimized for low prevalence.Digital ECGs with 12 leads or 15 leads (12-lead plus V3R, V4R and V7) were selected from multiple hospitals and separated into a testing and training database. Inclusion criterion was age less than 16 years. The reference for algorithm detection of VPE was cardiologist annotation of VPE for each ECG. The training database (n = 772) consisted of VPE ECGs (n = 37), normal ECGs (n = 492) and a high concentration of conduction defects, RBBB (n = 232) and LBBB (n = 11). The testing database was a random sample (n = 763). All ECGs were analyzed with the Philips DXL ECG Analysis algorithm for basic waveform measurements. Additional ECG features specific to VPE, mainly delta wave scoring, were calculated from the basic measurements and the average beat. A classifier based on decision tree bootstrap aggregation (tree bagger) was trained in multiple steps to select the number of decision trees and the 10 best features. The classifier accuracy was measured on the test database.The new algorithm detected pediatric VPE with a sensitivity of 78%, a specificity of 99.9%, a positive predictive value of 88% and negative predictive value of 99.7%.This new algorithm for detection of pediatric VPE performs well with a reasonable positive and negative predictive value despite the low prevalence in the general population.

Deep machine learning for the PANDA software trigger

Deep machine learning methods have been studied for the software trigger of the future PANDA experiment at FAIR, using Monte Carlo simulated data from the GEANT-based detector simulation framework PandaRoot. Ten physics channels that cover the main physics topics, including electromagnetic, exotic, charmonium, open charm, and baryonic reaction channels, have been investigated at four different anti-proton beam momenta. Different classification concepts and network architectures have been studied. Finally a residual convolutional neural network with four residual blocks in a binary classification scheme has been chosen due to its extendability, performance and stability. The presented study represents a feasibility study of a completely software-based trigger system. Compared to a conventional selection method, the deep machine learning approach achieved a significant efficiency gain of up to 200%, while keeping the background reduction factor at the required level of 1/1000. Furthermore, it is shown that the use of additional input variables can improve the data quality for subsequent analysis. This study shows that the PANDA software trigger can benefit greatly from the deep machine learning methods.

Robust fault diagnosis of quayside container crane gearbox based on 2D image representation in frequency domain and CNN

To accurately diagnose the quayside container crane (QCC) gearbox faults, this article proposes a method that combines the frequency-domain Markov transformation field (FDMTF) and multi-branch residual convolutional neural network (MBRCNN). Firstly, the gearbox vibration signal is converted into the frequency domain to reveal the components and amplitude of signals stably and concisely. Then, the one-dimensional frequency signal is encoded into the two-dimensional image by the Markov transformation field to capture the dynamic characteristics of signals. Thirdly, the MBRCNN network is constructed, which can extract multi-scale features and alleviate the problems caused by the deep network structure. Finally, the FDMTF image is fed into the constructed MBRCNN model for pattern recognition. The effectiveness of the proposed FDMTF–MBRCNN method is verified by two case studies. In Case 1, the diagnosis results of a benchmark dataset achieve 100% accuracy, better than seven state-of-the-art methods published in recent 3 years. In Case 2, the diagnosis results of the dataset collected from a 1:4 scaled test rig achieve 98.85% accuracy, better than eleven encoding methods and four convolutional neural network methods. It also can obtain a recognition accuracy of more than 94% under the conditions of small sample, different network hyper-parameters, or variable loads, which verifies its robustness. These case studies show that the FDMTF–MBRCNN method is expected to be applied to the actual fault diagnosis of QCC gearboxes.

RMA-CNN: A Residual Mixed-Domain Attention CNN for Bearings Fault Diagnosis and its Time-Frequency Domain Interpretability

Early fault diagnosis of bearings is crucial for ensuring safe and reliable operations. Convolutional Neural Networks (CNNs) have achieved significant breakthroughs in machinery fault diagnosis. However, complex and varying working conditions can lead to inter-class similarity and intra-class variability in datasets, making it more challenging for CNNs to learn discriminative features. Furthermore, CNNs are often considered "black boxes" and lack sufficient interpretability in the fault diagnosis field. To address these issues, this paper introduces a Residual Mixed-Domain AttentionCNN method, referred to as RMA-CNN. This method comprises multiple ResidualMixed Domain Attention Modules (RMAMs), each employing one attention mechanism to emphasize meaningful features in both time and channel domains. This significantly enhances the network's ability to learn fault-related features. Moreover, we conduct an in-depth analysis of the inherent feature learning mechanism of the attention module RMAM to improve the interpretability of CNNs in fault diagnosis applications. Experiments conducted on two datasets—a high-speed aeronautical bearing dataset and a motor bearing dataset—demonstrate that the RMA-CNN achieves remarkable results in diagnostic tasks.

A Lightweight Neural Network Based on GAF and ECA for Bearing Fault Diagnosis

A lightweight neural network fault diagnosis method based on Gramian angular field (GAF) feature map construction and efficient channel attention (ECA) optimization is presented herein to address the problem of the complex structure of traditional neural networks in bearing fault diagnosis. Firstly, a GAF is used to encode vibration signals into a temporal image. Secondly, the double-layer separation residual convolution neural network (DRCNN) is used to learn advanced features of the sample. The multi-branch structure is used as the receiving domain. ECA learns the correlation between feature channels. The extracted feature channels are adaptively weighted by adding a small additional computational cost. Finally, the method is tested and evaluated using wind turbine bearing data. The experimental results show that, compared with the traditional neural network, the DRCNN model based on GAF achieves higher diagnostic accuracy with less parameter calculation.