Multi-channel Feature Research Articles

A new deep transfer learning method for intelligent bridge damage diagnosis based on muti-channel sub-domain adaptation

Deep learning has been successfully and widely used in the field of intelligent diagnosis to accurately identify damage. These successes are based on the assumption that the training and test data are independent and identically distributed. They ignore that there is a distribution discrepancy between training (source domain) and test data (target domain) obtained under different working environments, loads, or bridges. This distribution discrepancy leads to severe performance degradation of the diagnostic model. To address these problems, a novel intelligent diagnosis method is proposed for bridges, namely, the multi-channel subdomain adaptation based deep transfer learning network. In this network, a multi-scale-based parallel multi-channel feature extractor and a multi-kernel local maximum mean discrepancy-based multi-channel subdomain adaptation module are proposed to learn domain-invariant features. Data sets from three different bridges are used for transfer diagnosis experiments to verify the generality and superiority of the method. This study will further facilitate the intelligentisation of structural health monitoring systems in the field of bridges.

Predicting future well performance for environmental remediation design using deep learning

In this study, we developed a deep learning (DL) framework with a multi-channel three-dimensional convolutional neural network (MC3D-CNN) to predict well performance and thereby assist future environmental remediation design. Such prediction of extraction well performance at designated locations is critical for configuring pump-and-treat (P&T) well network design and operation, setting reasonable target closure dates for overall remedying, and estimating remedy costs. The framework is developed with operational and monitoring data routinely collected during P&T remedy operations, including well extraction and injection rates as well as in situ contaminant concentrations. Traditionally, the collected data were rarely used for purposes other than assessing past well performance and the accuracy of the conceptual site model. However, recent advances in data-driven computational approaches enable better use of the large datasets to inform future well performance, enhance site characterization, and improve remediation planning. In this study, we established a DL framework to integrate transient three-dimensional contaminant plumes and multiple aquifer properties (e.g., hydraulic conductivity and hydrostratigraphic maps) to identify characteristic patterns controlling and representing extraction well mass recovery, aiming at providing future mass recovery estimates for existing wells and candidate wells at any proposed locations. We evaluated our framework by using a realistic synthetic dataset generated from a well-calibrated flow and transport model used in the 200 West Area of the U.S. Department of Energy’s Hanford Site in southeastern Washington state. The multi-channel feature in our framework allows integration of various types and temporal densities of training datasets for DL model development. Overall, we found that the trained DL model achieved an accuracy of over 90% in ranking extraction well performance in validation datasets, and over 80% in predicting high-performance-ranking well locations. This data-informed approach provides a flexible tool to support adaptive site management, streamline decision-making, and potentially reduce remediation time and costs. Our DL framework can be used as a filtering tool to improve the current P&T network optimization design by reducing the number of candidate well locations.

MCFN: A Multichannel Fusion Network for Sleep Apnea Syndrome Detection.

Sleep apnea syndrome (SAS) is the most common sleep disorder which affects human life and health. Many researchers use deep learning methods to automatically learn the features of physiological signals. However, these methods ignore the different effects of multichannel features from various physiological signals. To solve this problem, we propose a multichannel fusion network (MCFN), which learns the multilevel features through a convolution neural network on different respiratory signals and then reconstructs the relationship between feature channels with an attention mechanism. MCFN effectively fuses the multichannel features to improve the SAS detection performance. We conducted experiments on the Multi-Ethnic Study of Atherosclerosis (MESA) dataset, consisting of 2056 subjects. The experiment results show that our proposed network achieves an overall accuracy of 87.3%, which is better than other SAS detection methods and can better assist sleep experts in diagnosing sleep disorders.

A storm-time ionospheric TEC model with multichannel features by the spatiotemporal ConvLSTM network

A storm-time ionospheric TEC model with multichannel features by the spatiotemporal ConvLSTM network

FDA-MIMO Radar Target Recognition Based on SVM Classification with Multi-channel Feature Extraction

FDA-MIMO radar has good anti-active interference performance and has received extensive attention due to its distance dependence. However, there is still a lack of target and interference recognition methods for FDA-MIMO radar in complex environments. Therefore, in this paper, we propose a multi-channel feature extraction classification method based on the support vector machine (SVM), which realizes the classification of four active interferences and targets. In short, we classify the fine signal characteristics of different interference signals by extracting effective features and selecting efficient classifiers. Simulation results show that when the interference to noise ratio (INR) is greater than -15 dB, the recognition accuracy is greater than 0.95. It also shows that the proposed method can distinguish target and interference well in a variety of complex environments.

A Multi-Channel Feature Fusion CNN-Bi-LSTM Epilepsy EEG Classification and Prediction Model Based on Attention Mechanism

A Multi-Channel Feature Fusion CNN-Bi-LSTM Epilepsy EEG Classification and Prediction Model Based on Attention Mechanism

Concurrent Rate-Adaptive Reading With Passive RFIDs

Radio frequency identification (RFID)-assisted management systems have been widely applied in warehousing, logistics, retailing, etc. In these scenarios, RFID-aided applications, e.g., object tracking and human behavior sensing, rely on a high-efficiency tag reading to realize accurate analyses and timely responses. However, serious tag collisions in those large-scale RFID systems will inevitably lead to significant decreases in the tag reading rates. To meet the strict timeliness requirements of those practical applications, we aim to treat the individual reading rate for each item tag differently and focus more attention on those user-interactive ones. However, due to unpredictable user behaviors, it is impractical to infer the user-interactive tags in advance. In addition, keeping focusing on them for continuous monitoring despite user movements and multipath-prevalent environments is also challenging. To solve these problems, we propose Spotlight, the first concurrent rate-adaptive reading system in passive RFIDs. Spotlight screens the ID-agnostic user-interactive tags by proposing a multichannel feature for narrow-band RFID systems without any hardware or protocol modification and achieves rate-adaptive reading by implementing real-time MU-MIMO beamforming. Substantial experiments with 1000+ COTS RFID tags exhibit that Spotlight outperforms the commercial reader by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$2.7\times $ </tex-math></inline-formula> and the SDR-based reader by <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$6.12\times $ </tex-math></inline-formula> . In addition, Spotlight first proposes the online parallel decoding method to realize concurrency among multiple users, which breaks the commercial protocol’s throughput ceiling (37%) and achieves up to 59% throughputs.

Airport Capacity Prediction With Multisource Features: A Temporal Deep Learning Approach

Accurate airport capacity estimation is crucial for the secure and orderly operation of the aviation system. However, such estimation is a non-trivial task as capacity depends on various meteorological and operational features. The complex coupling characteristics among these multi-source features have proved to be challenging for most of the traditional regression models. Recently, enhanced by its excellent ability to mine nonlinear relationships, the machine learning methods trigger widely applications. However, due to the imbalance of features scatter and the neglect of temporal dependences in aviation systems, existing machine learning methods for airport capacity prediction still have room for improvement. In light of these, this paper presents a novel airport capacity prediction method based on the multi-channel fusion Transformer model (MF-Transformer). Besides the commonly used aviation features, we unprecedentedly harness the power of the high-dimensional meteorological feature for accurate prediction. As to the model, we construct a multi-channel feature fusion structure, which includes a three-channel network for multi-source features extraction and an attention-based feature fusion module between channels. In each channel, the Transformer-based model is utilized to capture the temporal dependences of features. We conduct experiments on the capacity prediction tasks of the Beijing Capital International Airport which is the largest airport in China and verify that the proposed MF-Transformer outperforms benchmarks under different prediction horizons.

A CNN-LSTM-PSO tool wear prediction method based on multi-channel feature fusion

In order to achieve predictive maintenance of CNC machining tools and to be able to change tools intelligently before tool wear is at a critical threshold, a CNN-LSTM tool wear prediction model based on particle swarm algorithm (PSO) optimization with multi-channel feature fusion is proposed. Firstly, the raw signals of seven channels of the machining process are collected using sensor technology and processed for noise reduction; secondly, the time-domain, frequency-domain and time-frequency-domain features of each channel signal are extracted, and a sample data set of spatio-temporal correlation of traffic flow is constructed by dimensionality reduction processing and information fusion of the above features; finally, the data set is input to the CNN-LSTM-PSO model for training and testing. The results show that the CNN-LSTM-PSO model can effectively predict tool wear with an average absolute error MAE value of 0.5848, a root mean square error RMSE value of 0.7281, and a coefficient of determination R2 value of 0.9964; and compared with the BP model, CNN model, LSTM model and CNN-LSTM model, its tool wear prediction accuracy improved by 7.56%, 2.60%, 2.98%, and 1.63%, respectively.

Strong Generalized Speech Emotion Recognition Based on Effective Data Augmentation

The absence of labeled samples limits the development of speech emotion recognition (SER). Data augmentation is an effective way to address sample sparsity. However, there is a lack of research on data augmentation algorithms in the field of SER. In this paper, the effectiveness of classical acoustic data augmentation methods in SER is analyzed, based on which a strong generalized speech emotion recognition model based on effective data augmentation is proposed. The model uses a multi-channel feature extractor consisting of multiple sub-networks to extract emotional representations. Different kinds of augmented data that can effectively improve SER performance are fed into the sub-networks, and the emotional representations are obtained by the weighted fusion of the output feature maps of each sub-network. And in order to make the model robust to unseen speakers, we employ adversarial training to generalize emotion representations. A discriminator is used to estimate the Wasserstein distance between the feature distributions of different speakers and to force the feature extractor to learn the speaker-invariant emotional representations by adversarial training. The simulation experimental results on the IEMOCAP corpus show that the performance of the proposed method is 2-9% ahead of the related SER algorithm, which proves the effectiveness of the proposed method.

Incipient fault diagnosis of analog circuit with ensemble HKELM based on fused multi-channel and multi-scale features

As an essential part in electronics-rich system, the failure of analog circuits will severely affect the system reliability and security. Incipient fault of analog circuit refers to the early stage of degradation fault where the fault characteristics are generally weak and almost indistinguishable. In order to enhance the reliability of electronic systems, it is necessary to diagnose incipient faults of analog circuits promptly and effectively. Existing approaches generally capture fault characteristics only from single signal, ignoring the valuable information inherent in different domains and scales. To address this problem, a novel diagnostic strategy based on multi-scale feature extraction and multi-channel feature fusion is designed to guarantee the completeness and richness of fault information. In this study, a deep extreme learning machine denoising auto-encoder (DELM-DAE) based method is proposed to conduct unsupervised multi-scale and multi-channel feature fusion to extract distinguishable features for incipient faults. The proposed method has higher learning efficiency and overcomes the common problem of low efficiency in deep learning model training. Meanwhile, in order to improve the ability to distinguish high-resolution features, an ensemble hybrid kernel extreme learning machine with novel roulette selection and weighted voting scheme is proposed to enhance the recognition performance and stability. In the verification experiment, the diagnosis accuracy on four typical circuits all reaches above 98%, which demonstrates that the proposed incipient fault diagnosis method for analog circuits has more conspicuous performance than other state-of-the-art methods.

The Msegram: A useful multichannel feature synchronous extraction tool for detecting rolling bearing faults

Multichannel signals collected by multiple sensors contain richer condition information of equipment than single-channel signals. However, such issues as simultaneous denoising, adaptive decomposition and synchronous extraction are still challenging for multichannel signals, which are beneficial to accurate fault diagnosis. Thus, a useful multichannel feature synchronous extraction tool is proposed for detecting rolling bearing faults, named as Msegram. First, a tensor synchronization denoising method based on high order singular value decomposition (HOSVD) is proposed for multichannel signal preprocessing. Original multichannel signals of testing bearings are constructed to be a third-order tensor by phase space reconstruction. Hereinto, a singular entropy increment is adopted to determine a reasonable singular order for each unfolding, and an optimal core tensor is obtained for local reconstruction analysis. Second, multi-layer K-value multivariate variational mode decomposition (MVMD) is designed after the multichannel noise reduction to realize synchronous adaptive filtering and decomposition for the multichannel signals. Third, inspired by the idea of the spectral kurtosis, a tower-shaped crest factor of envelope spectrum (EC) diagram similar to Fast Kurtogram (FK) is proposed to visualize the output of multichannel bearing fault feature results. According to the tower-shaped EC diagram with the maximum fault crest factor, the optimal analytic results of multichannel signals are selected and output to synchronously extract bearing fault features. Finally, repeatable simulations and two experimental fault cases of rolling bearings are implemented to demonstrate the practicability and effectiveness of the proposed method. The results show that the proposed method can successfully reveal the compound faults from experimental bearing and effectively identify the compound faults from locomotive wheelset bearing.

Multi-channel feature fusion networks with hard coordinate attention mechanism for maize disease identification under complex backgrounds

Maize leaf disease has a negative impact on food security. If maize leaf disease is not controlled in time, it can spread and affect corn quality and yield. Severe maize leaf disease may even result in no harvest at all. In the existing deep learning technology for maize leaf disease recognition, due to the complex growing environment of maize, it is easy to capture the complex background of maize in the image acquisition process, which brings difficulties to the recognition of maize leaf disease. Therefore, we propose a new network architecture, HCA-MFFNet, for maize leaf disease recognition in a complex background. In HCA-MFFNet, hard coordinated attention (HCA) assigned at different spatial scales was used to extract features from maize leaf disease images to reduce the influence of complex backgrounds. In order to retain the feature information of maize leaf disease in the sampling process to the maximum extent, a multi-feature fusion network that can extract the weight information in two spatial directions was constructed. The traditional convolutional layer was replaced by depthwise separable convolutional layer, reducing the number of parameters in the network. In order to verify the feasibility and effectiveness of this model in complex environment, we compare it with existing methods. The average recognition accuracy of the model is 97.75%, and the F1 value is 97.03%. The proposed method is better than the existing methods.

Recognition of Cervical Precancerous Lesions Based on Probability Distribution Feature Guidance.

Cervical cancer is a high incidence of cancer in women and cervical precancerous screening plays an important role in reducing the mortality rate. In this study, we proposed a multichannel feature extraction method based on the probability distribution features of the Acetowhite (AW) region to identify cervical precancerous lesions, with the overarching goal to improve the accuracy of cervical precancerous screening. A k-means clustering algorithm was first used to extract the cervical region images from the original colposcopy images. We then used a deep learning model called DeepLab V3+ to segment the AW region of the cervical image after the acetic acid experiment, from which the probability distribution map of the AW region after segmentation was obtained. This probability distribution map was fed into a neural network classification model for multichannel feature extraction, which resulted in the final classification performance. Results of the experimental evaluation showed that the proposed method achieved an average accuracy of 87.7%, an average sensitivity of 89.3%, and an average specificity of 85.6%. Compared with the methods that did not add segmented probability features, the proposed method increased the average accuracy rate, sensitivity, and specificity by 8.3%, 8%, and 8.4%, respectively. Overall, the proposed method holds great promise for enhancing the screening of cervical precancerous lesions in the clinic by providing the physician with more reliable screening results that might reduce their workload.

Health Status Assessment of Diesel Engine Valve Clearance Based on BFA-BOA-VMD Adaptive Noise Reduction and Multi-Channel Information Fusion.

Regarding the problem of the valve gap health status being difficult to assess due to the complex composition of the condition monitoring signal during the operation of the diesel engine, this paper proposes an adaptive noise reduction and multi-channel information fusion method for the health status assessment of diesel engine valve clearance. For the problem of missing fault information of single-channel sensors in condition monitoring, we built a diesel engine valve clearance preset simulation test bench and constructed a multi-sensor acquisition system to realize the acquisition of diesel engine multi-dimensional cylinder head signals. At the same time, for the problem of poor adaptability of most signal analysis methods, the improved butterfly optimization algorithm by the bacterial foraging algorithm was adopted to adaptively optimize the key parameter for variational mode decomposition, with discrete entropy as the fitness value. Then, to reduce the uncertainty of artificially selecting fault characteristics, the characteristic parameters with a higher recognition degree of diesel engine signal were selected through characteristic sensitivity analysis. To achieve an effective dimensionality reduction integration of multi-channel features, a stacked sparse autoencoder was used to achieve deep fusion of the multi-dimensional feature values. Finally, the feature samples were entered into the constructed one-dimensional convolutional neural network with a four-layer parameter space for training to realize the health status assessment of the diesel engine. In addition, we verified the effectiveness of the method by carrying out valve degradation simulation experiments on the diesel engine test bench. Experimental results show that, compared with other common evaluation methods, the method used in this paper has a better health state evaluation effect.

Improving Node Classification through Convolutional Networks Built on Enhanced Message-Passing Graph.

Enhancing message propagation is critical for solving the problem of node classification in sparse graph with few labels. The recently popularized Graph Convolutional Network (GCN) lacks the ability to propagate messages effectively to distant nodes because of over-smoothing. Besides, the GCN with numerous trainable parameters suffers from overfitting when the labeled nodes are scarce. This article addresses the problem via building GCN on Enhanced Message-Passing Graph (EMPG). The key idea is that node classification can benefit from various variants of the input graph that can propagate messages more efficiently, based on the assumption that the structure of each variant is reasonable when more unlabeled nodes are labeled properly. Specifically, the proposed method first maps the nodes to a latent space through graph embedding that captures the structural information of the input graph. Considering the node attributes together, the proposed method constructs the EMPG by adding connections between the nodes in close proximity in the latent space. With the help of the added connections, the EMPG allows a node to propagate its message to the right nodes at long distances, so that the GCN built on the EMPG need not stack multiple layers. As a result, over-smoothing is avoided. However, dense connections may cause message propagation saturation and lead to overfitting. Seeing the EMPG as an accumulation of some potential variants of the original graph, the proposed method utilizes dropout to extract a group of variants from the EMPG and then builds multichannel GCNs on them. The multichannel features learned from different dropout EMPGs are aggregated to compute the final prediction jointly. The proposed method is flexible, as a brod range of GCNs can be incorporated easily. Additionally, it is efficient and robust. Experimental results demonstrate that the proposed method yields improvements in node classification.

Self-Sparse Generative Adversarial Networks

Generative Adversarial Networks (GANs) are an unsupervised generative model that learns data distribution through adversarial training. However, recent experiments indicated that GANs are difficult to train due to the requirement of optimization in the high dimensional parameter space and the zero gradient problem. In this work, we propose a Self Sparse Generative Adversarial Network (Self-Sparse GAN) that reduces the parameter space and alleviates the zero gradient problem. In the Self-Sparse GAN, we design a Self-Adaptive Sparse Transform Module (SASTM) comprising the sparsity decomposition and feature-map recombination, which can be applied on multi-channel feature maps to obtain sparse feature maps. The key idea of Self-Sparse GAN is to add the SASTM following every deconvolution layer in the generator, which can adaptively reduce the parameter space by utilizing the sparsity in multi-channel feature maps. We theoretically prove that the SASTM can not only reduce the search space of the convolution kernel weight of the generator but also alleviate the zero gradient problem by maintaining meaningful features in the Batch Normalization layer and driving the weight of deconvolution layers away from being negative. The experimental results show that our method achieves the best FID scores for image generation compared with WGAN-GP on MNIST, Fashion-MNIST, CIFAR-10, STL-10, mini-ImageNet, CELEBA-HQ, and LSUN bedrooms, and the relative decrease of FID is 4.76% ~ 21.84%.

Rendered Image Superresolution Reconstruction with Multichannel Feature Network

In the process of film and television production, clear images can give the audience a real sensory experience, but high-resolution images require a massive amount of production time and highly specialized imaging equipment, which is not a cost-effective solution at the moment. To achieve a better cost efficiency during video production, we propose a multichannel featured superresolution network model that utilizes rendered low-resolution images according to their characteristics. This model includes a feature extraction layer, a series of subnetworks, and a reconstruction module. Inside the network model, a series of subnetworks are cascaded to improve the information flow from coarse to fine, which helps to fully extract the depth, normal vector, edge, and texture features from low-resolution rendered images to reconstruct the high-resolution image. Additionally, residual learning is introduced at each stage to further improve the reconstruction performance. We experiment with the model on the classic Disney Monte Carlo datasets and compare it with several related algorithms. The results show that our algorithm is able to reconstruct the image with clearer details and texture. Thus, our research not only helps to maintain the audience’s sensory experience but also increases the efficiency of film and television production, which also brings considerable economic benefits.

Side-Scan Sonar Image Segmentation Based on Multi-Channel CNN for AUV Navigation.

The AUV (Autonomous Underwater Vehicle) navigation process relies on the interaction of a variety of sensors. The side-scan sonar can collect underwater images and obtain semantic underwater environment information after processing, which will help improve the ability of AUV autonomous navigation. However, there is no practical method to utilize the semantic information of side scan sonar image. A new convolutional neural network model is proposed to solve this problem in this paper. The model is a standard codec structure, which extracts multi-channel features from the input image and then fuses them to reduce parameters and strengthen the weight of feature channels. Then, a larger convolution kernel is used to extract the features of large-scale sonar images more effectively. Finally, a parallel compensation link with a small-scale convolution kernel is added and spliced with features extracted from a large convolution kernel in the decoding part to obtain features of different scales. We use this model to conduct experiments on self-collected sonar data sets, which were uploaded on github. The experimental results show that ACC and MIoU reach 0.87 and 0.71, better than other classical small-order semantic segmentation networks. Furthermore, the 347.52 g FOLP and the number of parameters around 13 m also ensure the computing speed and portability of the network. The result can extract the semantic information of the side-scan sonar image and assist with AUV autonomous navigation and mapping.

Tied Bilateral learning for Aquaculture Image Enhancement

Aquaculture plays a critical role in food security and nutrition strategies. The application of intelligent aquaculture technology has shown promising performance in improving aquaculture productivity and increasing economic benefits with its rapid advancement and good prospects. However, degraded underwater images have hampered the existing computer vision applications in intelligent aquaculture. To this end, a novel Tied Bilateral learning network is proposed for Aquaculture Image Enhancement (TBAIE), which improves the degraded aquaculture images to meet the requirements of various computer vision applications in aquaculture. Concretely, a novel multiple tied guidance module is designed to generate a multi-channel feature map and capture long-range features based on input. Then, a feature fusion module is introduced with a novel tied attention block to blend the feature and suppress noise with a low computational resource. Experimental results demonstrate that the proposed TBAIE can improve the quality of aquaculture images and remove color distortion. Moreover, TBAIE can achieve state-of-the-art in quantitative and qualitative metrics and meet the practical requirements of different aquaculture vision tasks.