Convolutional Neural Network Transfer Research Articles

Deep Reconstruction Transfer Convolutional Neural Network for Rolling Bearing Fault Diagnosis.

Deep transfer learning has been widely used to improve the versatility of models. In the problem of cross-domain fault diagnosis in rolling bearings, most models require that the given data have a similar distribution, which limits the diagnostic effect and generalization of the model. This paper proposes a deep reconstruction transfer convolutional neural network (DRTCNN), which satisfies the domain adaptability of the model under cross-domain conditions. Firstly, the model uses a deep reconstruction convolutional automatic encoder for feature extraction and data reconstruction. Through sharing parameters and unsupervised training, the structural information of target domain samples is effectively used to extract domain-invariant features. Secondly, a new subdomain alignment loss function is introduced to align the subdomain distribution of the source domain and the target domain, which can improve the classification accuracy by reducing the intra-class distance and increasing the inter-class distance. In addition, a label smoothing algorithm considering the credibility of the sample is introduced to train the model classifier to avoid the impact of wrong labels on the training process. Three datasets are used to verify the versatility of the model, and the results show that the model has a high accuracy and stability.

Ensemble of Deep Features for Breast Cancer Histopathological Image Classification

Abstract Analysis of histopathological images (HIs) is crucial for detecting breast cancer (BR). However, because they vary, it is still very difficult to extract well-designed elements. Deep learning (DL) is a recent development that is used to extract high-level features. However, DL techniques continue to confront several difficult problems, such as the need for sufficient training data for DL models, which reduces the classification findings. In this study, an ensemble deep transfer convolutional neural network is presented to address this problem. The pre-trained models (ResNet50 and MobileNet) are employed to extract high-level features by freezing the front layer parameters while fine-tuning the last layers. In the proposed ensemble framework, KNN, SVM, logistic regression and neural networks are used as base classifiers. The majority vote and product approaches are used to integrate the predictions of each separate classifier. In the benchmark BreaKHis dataset, the suggested ensemble model is compared to some current approaches. It demonstrates that while the ensemble model obtains a considerable accuracy of 97.72% for the multiclass classification test, it achieves an accuracy of 99.2% for the binary task. The suggested ensemble model’s effectiveness in extracting useful features for BR images is demonstrated by comparison with existing cutting-edge models.

High-Impedance Fault Detection Methodology Using Time–Frequency Spectrum and Transfer Convolutional Neural Network in Distribution Networks

High-impedance fault (HIF) detection has always been difficult in distribution networks due to the lack of field data and the large difference between field and simulation waveforms. Based on the characteristics of zero-sequence currents, a novel HIF detection methodology is proposed, which combines time–frequency spectrum (TFS) and transfer convolutional neural network (TCNN). First, the TFSs are acquired by applying continuous wavelet transform (CWT) to the collected zero-sequence currents. Then, the TFSs of simulated zero-sequence currents are utilized for training source-domain convolutional neural network (SCNN). Next, the SCNN is transfer learned with very few TFSs of field zero-sequence currents to obtain TCNN. The performance of the proposed method is verified by simulation samples and field samples. The results show that the proposed method can effectively extract fault features from small-scale training samples under different fault circumstances. Besides, TCNN can adaptively extract the effective features of field HIF and detect field HIF more accurately than SCNN. Finally, this article provides a visualization scheme for interpretability of the neural network, which offers visual explanations for the decision-making basis of the neural network.

Joint Diagnostic Method of Tumor Tissue Based on Hyperspectral Spectral-Spatial Transfer Features

In order to improve the clinical application of hyperspectral technology in the pathological diagnosis of tumor tissue, a joint diagnostic method based on spectral-spatial transfer features was established by simulating the actual clinical diagnosis process and combining micro-hyperspectral imaging with large-scale pathological data. In view of the limited sample volume of medical hyperspectral data, a multi-data transfer model pre-trained on conventional pathology datasets was applied to the classification task of micro-hyperspectral images, to explore the differences in spectral-spatial transfer features in the wavelength of 410–900 nm between tumor tissues and normal tissues. The experimental results show that the spectral-spatial transfer convolutional neural network (SST-CNN) achieved a classification accuracy of 95.46% for the gastric cancer dataset and 95.89% for the thyroid cancer dataset, thus outperforming models trained on single conventional digital pathology and single hyperspectral data. The joint diagnostic method established based on SST-CNN can complete the interpretation of a section of data in 3 min, thus providing a new technical solution for the rapid diagnosis of pathology. This study also explored problems involving the correlation between tumor tissues and typical spectral-spatial features, as well as the efficient transformation of conventional pathological and transfer spectral-spatial features, which solidified the theoretical research on hyperspectral pathological diagnosis.

Data-driven lithium-ion batteries capacity estimation based on deep transfer learning using partial segment of charging/discharging data

Accurate estimation of lithium-ion battery capacity is crucial for ensuring its safety and reliability. While data-driven modelling is a common approach for capacity estimation, obtaining cycling data during charging/discharging processes can be challenging. Collecting cycling data under various charging/discharging protocols is often unrealistic, and the collected data can be fragmented due to the random nature of working conditions in practice. To address these issues, we propose a deep transfer learning method that uses partial segments of charging/discharging data for battery capacity estimation. The proposed method utilizes capacity increment features of partial charging/discharging segments that is designed to satisfy practical scenarios. A deep transfer convolutional neural network (DTCNN) is trained with both source and target data, and a fine-tuning strategy is employed to effectively eliminate distribution discrepancies between different battery types or charging/discharging protocols, leading the improved estimation accuracy. Experimental results demonstrate that the proposed method accurately estimates the lithium-ion battery capacity, with values of RMSE, MAPE, and MD-MAPE of only 0.0220, 0.0247, and 0.0194, respectively, when using partial segments. These results highlight the promising prospects of the proposed method for lithium-ion battery capacity estimation.

Intelligent machine fault diagnosis based on deep transfer convolutional neural network and extreme learning machine

With the in-depth study of researchers in the field of fault diagnosis, many machine fault diagnosis methods based on deep learning have been proposed. These methods have achieved remarkable results, but some practical issues still should be solved, such as lack of sufficient training data with labels and long training time. In this research, a machine fault diagnosis method using deep transfer convolutional neural network (DTCNN) and extreme learning machine (ELM) is proposed. Firstly, continuous wavelet transform (CWT) is adopted to transform vibration signals into 2D time-frequency images. Then, the optimal DTCNN pre-trained by ImageNet dataset is selected to extract high-level features of time-frequency images. The extracted high-level features further are input to the ELM classifier for fault classification. Finally, the extracted high-level features further are input to the ELM classifier for fault classification. The effectiveness and efficiency of the proposed method are verified on two well-known datasets, including the Case Western Reserve University (CWRU) motor bearing dataset and the KAt bearing dataset of Paderborn University. The experimental results show that the proposed method can greatly reduce the computational time of the model while ensuring high accuracy of diagnosis, and DTCNN-ELM outperforms other state-of-the-art methods.

Fault diagnosis of rolling bearing based on online transfer convolutional neural network

In order to achieve online fault diagnosis of rolling bearing effectively, this paper proposes a rolling bearing fault diagnosis model based on online transfer convolutional neural network (OTCNN). Firstly, offline convolutional neural network (Off-CNN) and online convolutional neural network (On-CNN) with the same model structure are constructed, and multi-channel data fusion and gray image conversion are used as the input of the model. Then, the source domain features in the fully connected layer and the model parameters are obtained by the pre-trained Off-CNN. Finally, the parameters of the On-CNN are initialized by the parameters of the Off-CNN, and the pre-trained source domain features can be used to achieve domain adaptation. A comprehensive analysis with the traditional algorithms is also performed, the results demonstrate that the proposed model can reduce the training time by half while ensuring the accuracy of it.

Balanced Medical Image Classification with Transfer Learning and Convolutional Neural Networks

This paper aims to propose a tool for image classification in medical diagnosis decision support, in a context where computational power is limited and then specific, high-speed computing infrastructures cannot be used (mainly for economic and energy consuming reasons). The proposed method combines a deep neural networks algorithm with medical imaging procedures and is implemented to allow an efficient use on affordable hardware. The convolutional neural network (CNN) procedure used VGG16 as its base architecture, using the transfer learning technique with the parameters obtained in the ImageNet competition. Two convolutional blocks and one dense block were added to this architecture. The tool was developed and calibrated on the basis of five common lung diseases using 5430 images from two public datasets and the transfer learning technique. The holdout ratios of 90% and 10% for training and testing, respectively, were obtained, and the regularization tools were dropout, early stopping, and Lasso regularization (L2). An accuracy (ACC) of 56% and an area under the receiver-operating characteristic curve (ROC—AUC) of 50% were reached in testing, which are suitable for decision support in a resource-constrained environment.

A deep reinforcement transfer convolutional neural network for rolling bearing fault diagnosis

Deep neural networks highly depend on substantial labeled samples when identifying bearing fault. However, in some practical situations, it is very difficult to collect sufficient labeled samples, which limits the application of deep neural networks in practical engineering. Therefore, how to use limited labeled samples to complete fault diagnosis tasks is an urgent problem. In this paper, a deep reinforcement transfer convolutional neural network (DRTCNN) is developed to tackle the problem. Firstly, an intelligent diagnosis agent constructed by a convolutional neural network is trained to obtain maximum long-term cumulative rewards, which is characterized by the ability to autonomously learn the latent relationship between fault samples and corresponding labels. Secondly, the parameter transfer learning method is utilized to establish a target task agent of DRTCNN. Finally, limited labeled target domain fault samples and the training mechanism of deep Q-network are employed to train the target task agent for performing target diagnosis tasks. Two diagnosis cases are conducted to verify the effectiveness of the proposed method when only limited labeled target domain fault samples are available.

Noise Annoyance Prediction of Urban Substation Based on Transfer Learning and Convolutional Neural Network

The noise pollution caused by urban substations is an increasingly serious problem, as is the issue of local residents being disturbed by substation noise. To accurately assess the degree of noise annoyance caused by substations to surrounding residents, we established a noise annoyance prediction model based on transfer learning and a convolution neural network. Using the model, we took the noise spectrum as the input, the subjective evaluation result as the target output, and the AlexNet network model with a modified output layer and corresponding parameters as the pre-training model. In a fixed learning rate and epoch setting, the influence of different mini-batch size values on the prediction accuracy of the model was compared and analyzed. The results showed that when the mini-batch size was set to 4, 8, 16, and 32, all the data sets had convergence after 90 iterations. The root mean square error (RMSE) of all validation sets was lower than 0.355, and the loss of all validation sets was lower than 0.067. As the mini-batch size increased, the RMSE, loss, and mean absolute error (MAE) of the verification set gradually increased, while the number of iterations and the training duration decreased gradually. In this test, a mini-batch size value of four was appropriate. The resultant convolutional neural network model showed high accuracy and robustness, and the error between the prediction result and the subjective evaluation result was between 2% and 7%. The model comprehensively reflects the objective metrics affecting subjective perception, and accurately describes the subjective perception of urban substation noise on human ears.

High-cycle fatigue S-N curve prediction of steels based on a transfer learning-guided convolutional neural network

The evaluation and prediction of fatigue properties are crucial for metallic materials. Although the determination of S-N curves represents the most important methods for evaluating such properties, its fatigue testing is costly and time-consuming. Furthermore, fatigue testing involves different test conditions, thereby complicating the evaluation of the fatigue properties. This study develops a transfer convolutional neural network (TR-CNN) framework, in which the prediction of the reversed torsion S-N curves of steels is transferred from rotating bending S-N curves. In the TR-CNN framework, the source CNN models for rotating-bending curve prediction are first trained based on the composition and process conditions. Subsequently, based on the source models, the reversed torsion S-N curves are estimated by training the TR-CNN models based on only a small dataset. After proving the rationality of the framework, its universality with respect to different amounts of data is further investigated. The reversed torsion curves under small-sample conditions (22 samples) are predicted accurately by the TR-CNN. Additionally, the TR-CNN models remain accurate under varying amounts of data (22-112 samples), showing excellent generality for different amounts of fatigue data. The predictive capability of the TR-CNN models is improved by introducing tensile properties into the source models. The proposed TR-CNN framework can significantly reduce the cost of evaluating fatigue properties, and the prediction of S-N curves can be optimized by combining the transfer framework and low-cost properties related to fatigue.

YOLO-TN: ultrasound diagnosis of thyroid nodules based on transfer learning and convolutional neural networks

YOLO-TN: ultrasound diagnosis of thyroid nodules based on transfer learning and convolutional neural networks

Maize leaf disease identification using deep transfer convolutional neural networks

Gray leaf spot, common rust, and northern leaf blight are three common maize leaf diseases that cause great economic losses to the worldwide maize industry. Timely and accurate disease identification can reduce economic losses, pesticide usage, and ensure maize yield and food security. Deep learning methods, represented by convolutional neural networks (CNNs), provide accurate, effective, and automatic diagnosis on server platforms when enormous training data is available. Restricted by dataset scale and application scenarios, CNNs are difficult to identify small-scale data sets on mobile terminals, while the lightweight networks, designed for the mobile terminal, achieve a better balance between efficiency and accuracy. This paper proposes a two-staged deep-transfer learning method to identify maize leaf diseases in the field. During the deep learning period, 8 deep and 4 lightweight CNN models were trained and compared on the Plant Village dataset, and ResNet and MobileNet achieved test accuracy of 99.48% and 98.69% respectively, which were then migrated onto the field maize leave disease dataset collected on mobile phones. By using layer-freezing and fine-tuning strategies on ResNet and MobileNet, fine-tuned MobileNet achieved the best accuracy of 99.11%. Results confirmed that disease identification performance from lightweight CNNs was not inferior to that of deep CNNs and transfer learning training efficiency was higher when lacking training samples. Besides, the smaller gaps between source and target domains, the better the identification performance for transfer learning. This study provides an application example for maize disease identification in the field using deep-transfer learning and provides a theoretical basis for intelligent maize leaf disease identification from images captured with mobile devices. Keywords: maize leaf disease, deep learning, transfer learning, convolutional neural networks DOI: 10.25165/j.ijabe.20221505.6658 Citation: Ma Z, Wang Y, Zhang T S, Wang H G, Jia Y J, Gao R, et al. Maize leaf disease identification using deep transfer convolutional neural networks. Int J Agric & Biol Eng, 2022; 15(5): 187–195.

Cross‐machine intelligent fault diagnosis of gearbox based on deep learning and parameter transfer

With the rapid development of artificial intelligence technologies, data-driven methods have significantly contributed to the intelligent monitoring and diagnosis of mechanical systems. However, the state-of-the-art approaches, especially the deep learning-based ones, implicitly assume the availability of large amounts of labeled fault data for supervised training, which is often infeasible due to the highly reliable system design in the field. In this research, a deep transfer convolutional neural network (CNN) scheme is proposed to enhance the diagnosis performance when dealing with insufficient training data in the target domain. By utilizing transfer learning, rich but relevant feature representation can be learnt from massive data in the source domain. The learnt weights and biases in the source domain are transferred to the target task as the initial parameter values. Then, the transferred parameters are properly fine-tuned with the small labeled datasets in the target domain. To avoid overfitting in the case of scarcely labeled samples in the target domain, global average pooling (GAP) is introduced to replace the fully-connected layers, and the traditional architecture in CNN is modified, to reduce the number of trainable parameters. Finally, by fully considering the transfer scenarios between diverse operating conditions and diverse machines, the cross-machine transfer experiments are designed with three gearbox datasets provided by the Prognostic and Health Management (PHM) 2009 conference, the Tsinghua University, and the University of Alberta. The results demonstrate the effectiveness of the proposed method with scarce labeled samples in the target domain.

Intelligent classification of ground-based visible cloud images using a transfer convolutional neural network and fine-tuning

Here a classification method for ground-based visible images is proposed based on a transfer convolutional neural network (TCNN). This approach combines the ability of deep learning (DL) and transfer learning (TL). A sample database containing all ten cloud types was used; this database was expanded four-fold using enhancement processing. AlexNet was chosen as the basic convolutional neural network (CNN), with the ImageNet database being used for pre-transfer. The optimal method, once determined by layer-by-layer fine-tuning, was used to test the classification effects for ten cloud types. The proposed method achieved 92.3% recognition accuracy for all ten ground-based cloud types.

An interpretable deep learning approach for calibration transfer among multiple near-infrared instruments

Spectroscopic techniques have been widely applied in agricultural applications. The development of calibration transfer is promising for the robust analysis of spectral data collected by varying instruments. The reliance on standard samples for standardization approaches remains a critical challenge for on-site applications. In this study, a deep learning approach, named DeepTranSpectra (DTS), is proposed to transfer convolutional neural network models among multiple near-infrared spectrometers with different types. The proposed DTS approach effectively avoids the requirements for standard samples by using labeled samples of slave instruments. The calibration transfer analysis is investigated on a soybean meal and one wheat dataset for predicting moisture and crude protein contents. The developed DTS approach demonstrates improved transfer performance compared with three popular standardization approaches, including piecewise direct standardization (PDS), canonical correlation analysis (CCA), and slope and bias correction (SBC). A feature visualization method is leveraged to interpret the transfer mechanism of the DTS approach. The interpretation results show that the DTS approach refines the model parameters to adapt to slave devices based on critical features of the master calibration. The DTS approach provides advanced reliability under different sample selections in Monte-Carlo cross-validation. The integration of deep learning approaches with calibration transfer analysis facilitates agricultural applications for emerging deep learning-based chemometric analysis.

Deep Learning for Identification of Acute Illness and Facial Cues of Illness.

Background: The inclusion of facial and bodily cues (clinical gestalt) in machine learning (ML) models improves the assessment of patients' health status, as shown in genetic syndromes and acute coronary syndrome. It is unknown if the inclusion of clinical gestalt improves ML-based classification of acutely ill patients. As in previous research in ML analysis of medical images, simulated or augmented data may be used to assess the usability of clinical gestalt.Objective: To assess whether a deep learning algorithm trained on a dataset of simulated and augmented facial photographs reflecting acutely ill patients can distinguish between healthy and LPS-infused, acutely ill individuals.Methods: Photographs from twenty-six volunteers whose facial features were manipulated to resemble a state of acute illness were used to extract features of illness and generate a synthetic dataset of acutely ill photographs, using a neural transfer convolutional neural network (NT-CNN) for data augmentation. Then, four distinct CNNs were trained on different parts of the facial photographs and concatenated into one final, stacked CNN which classified individuals as healthy or acutely ill. Finally, the stacked CNN was validated in an external dataset of volunteers injected with lipopolysaccharide (LPS).Results: In the external validation set, the four individual feature models distinguished acutely ill patients with sensitivities ranging from 10.5% (95% CI, 1.3–33.1% for the skin model) to 89.4% (66.9–98.7%, for the nose model). Specificity ranged from 42.1% (20.3–66.5%) for the nose model and 94.7% (73.9–99.9%) for skin. The stacked model combining all four facial features achieved an area under the receiver characteristic operating curve (AUROC) of 0.67 (0.62–0.71) and distinguished acutely ill patients with a sensitivity of 100% (82.35–100.00%) and specificity of 42.11% (20.25–66.50%).Conclusion: A deep learning algorithm trained on a synthetic, augmented dataset of facial photographs distinguished between healthy and simulated acutely ill individuals, demonstrating that synthetically generated data can be used to develop algorithms for health conditions in which large datasets are difficult to obtain. These results support the potential of facial feature analysis algorithms to support the diagnosis of acute illness.

A novel adversarial transfer learning in deep convolutional neural network for intelligent diagnosis of gas‐insulated switchgear insulation defect

Recently, numerous data-driven fault diagnosis methods have been developed, and the tasks involving the same distribution of training and test data have been well solved. However, considering the particularity of gas-insulated switchgear (GIS), collecting massive data, especially with the same distribution, is difficult. Therefore, existing fault diagnosis methods hardly achieve satisfactory insulation defect diagnosis with small datasets. Aiming at solving this problem, a novel domain adversarial transfer convolutional neural network (DATCNN) is proposed, realising the diagnosis of GIS insulation defects on small samples. First, a residual CNN is built to learn feature representations from the source and target domains. Second, the domain adversarial training strategy is used for feature transfer, where a conditional adversarial mechanism is introduced, and the joint distribution of features and labels is improved to a random linear combination, which realises the simultaneous adaptation of features and labels. Finally, the Nesterov accelerated gradient descent optimisation algorithm is used to speed up the gradient convergence. DATCNN has 99.15% and ≥89.5% diagnosis accuracy for GIS insulation defects in the laboratory and on-site, respectively. Comprehensive experiment results show the effectiveness and superiority of the proposed method in diagnosing GIS insulation defects with small samples.

Lightweight transfer nets and adversarial data augmentation for photovoltaic series arc fault detection with limited fault data

Incidents of DC series arc faults in Photovoltaic (PV) systems are becoming more common, posing significant threat to properties and human safety. Machine Learning (ML) based methods, developed recently, have demonstrated better performance in many fault diagnosis tasks. However, an unresolved challenge affecting their performance is the problem caused by difference between the source domain data used during the development and the target domain data encountered in operation in the field. Furthermore, the fault data in the target-domain are usually rare or not available for model training. Another constraint is that complex models are difficult to operate in real-time. This paper proposes a cross-domain DC series arc fault detection framework based on Lightweight Transfer Convolutional Neural Networks with Adversarial Data Augmentation (LTCNN-ADA) using limited target-domain fault data. Four datasets are prepared using different power sources and inverters in different operating conditions. The proposed framework is validated through comprehensive studies and experiments with different amount of fault data.

Intelligent Fault Diagnosis of Rotor-Bearing System Under Varying Working Conditions With Modified Transfer Convolutional Neural Network and Thermal Images

The existing intelligent fault diagnosis methods of rotor-bearing system mainly focus on vibration analysis under steady operation, which has low adaptability to new scenes. In this article, a new framework for rotor-bearing system fault diagnosis under varying working conditions is proposed by using modified convolutional neural network (CNN) with transfer learning. First, infrared thermal images are collected and used to characterize the health condition of rotor-bearing system. Second, modified CNN is developed by introducing stochastic pooling and Leaky rectified linear unit to overcome the training problems in classical CNN. Finally, parameter transfer is used to enable the source modified CNN to adapt to the target domain, which solves the problem of limited available training data in the target domain. The proposed method is applied to analyze thermal images of rotor-bearing system collected under different working conditions. The results show that the proposed method outperforms other cutting edge methods in fault diagnosis of rotor-bearing system.