Traditional data-driven fault diagnosis methods depend on expert experience to manually extract effective fault features of signals,which has certain limitations.Conversely,deep learning techniques have gained promine...Traditional data-driven fault diagnosis methods depend on expert experience to manually extract effective fault features of signals,which has certain limitations.Conversely,deep learning techniques have gained prominence as a central focus of research in the field of fault diagnosis by strong fault feature extraction ability and end-to-end fault diagnosis efficiency.Recently,utilizing the respective advantages of convolution neural network(CNN)and Transformer in local and global feature extraction,research on cooperating the two have demonstrated promise in the field of fault diagnosis.However,the cross-channel convolution mechanism in CNN and the self-attention calculations in Transformer contribute to excessive complexity in the cooperative model.This complexity results in high computational costs and limited industrial applicability.To tackle the above challenges,this paper proposes a lightweight CNN-Transformer named as SEFormer for rotating machinery fault diagnosis.First,a separable multiscale depthwise convolution block is designed to extract and integrate multiscale feature information from different channel dimensions of vibration signals.Then,an efficient self-attention block is developed to capture critical fine-grained features of the signal from a global perspective.Finally,experimental results on the planetary gearbox dataset and themotor roller bearing dataset prove that the proposed framework can balance the advantages of robustness,generalization and lightweight compared to recent state-of-the-art fault diagnosis models based on CNN and Transformer.This study presents a feasible strategy for developing a lightweight rotating machinery fault diagnosis framework aimed at economical deployment.展开更多
Recently,video-based fire detection technology has become an important research topic in the field of machine vision.This paper proposes a method of combining the classification model and target detection model in dee...Recently,video-based fire detection technology has become an important research topic in the field of machine vision.This paper proposes a method of combining the classification model and target detection model in deep learning for fire detection.Firstly,the depthwise separable convolution is used to classify fire images,which saves a lot of detection time under the premise of ensuring detection accuracy.Secondly,You Only Look Once version 3(YOLOv3)target regression function is used to output the fire position information for the images whose classification result is fire,which avoids the problem that the accuracy of detection cannot be guaranteed by using YOLOv3 for target classification and position regression.At the same time,the detection time of target regression for images without fire is greatly reduced saved.The experiments were tested using a network public database.The detection accuracy reached 98%and the detection rate reached 38fps.This method not only saves the workload of manually extracting flame characteristics,reduces the calculation cost,and reduces the amount of parameters,but also improves the detection accuracy and detection rate.展开更多
This letter deals with the frequency domain Blind Source Separation of Convolutive Mixtures (CMBSS). From the frequency representation of the "overlap and save", a Weighted General Discrete Fourier Transform...This letter deals with the frequency domain Blind Source Separation of Convolutive Mixtures (CMBSS). From the frequency representation of the "overlap and save", a Weighted General Discrete Fourier Transform (WGDFT) is derived to replace the traditional Discrete Fourier Transform (DFT). The mixing matrix on each frequency bin could be estimated more precisely from WGDFT coefficients than from DFT coefficients, which improves separation performance. Simulation results verify the validity of WGDFT for frequency domain blind source separation of convolutive mixtures.展开更多
Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware reso...Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.展开更多
Background The use of remote photoplethysmography(rPPG)to estimate blood volume pulse in a noncontact manner has been an active research topic in recent years.Existing methods are primarily based on a singlescale regi...Background The use of remote photoplethysmography(rPPG)to estimate blood volume pulse in a noncontact manner has been an active research topic in recent years.Existing methods are primarily based on a singlescale region of interest(ROI).However,some noise signals that are not easily separated in a single-scale space can be easily separated in a multi-scale space.Also,existing spatiotemporal networks mainly focus on local spatiotemporal information and do not emphasize temporal information,which is crucial in pulse extraction problems,resulting in insufficient spatiotemporal feature modelling.Methods Here,we propose a multi-scale facial video pulse extraction network based on separable spatiotemporal convolution(SSTC)and dimension separable attention(DSAT).First,to solve the problem of a single-scale ROI,we constructed a multi-scale feature space for initial signal separation.Second,SSTC and DSAT were designed for efficient spatiotemporal correlation modeling,which increased the information interaction between the long-span time and space dimensions;this placed more emphasis on temporal features.Results The signal-to-noise ratio(SNR)of the proposed network reached 9.58dB on the PURE dataset and 6.77dB on the UBFC-rPPG dataset,outperforming state-of-the-art algorithms.Conclusions The results showed that fusing multi-scale signals yielded better results than methods based on only single-scale signals.The proposed SSTC and dimension-separable attention mechanism will contribute to more accurate pulse signal extraction.展开更多
Most of the existing algorithms for blind sources separation have a limitation that sources are statistically independent. However, in many practical applications, the source signals are non- negative and mutual stati...Most of the existing algorithms for blind sources separation have a limitation that sources are statistically independent. However, in many practical applications, the source signals are non- negative and mutual statistically dependent signals. When the observations are nonnegative linear combinations of nonnegative sources, the correlation coefficients of the observations are larger than these of source signals. In this letter, a novel Nonnegative Matrix Factorization (NMF) algorithm with least correlated component constraints to blind separation of convolutive mixed sources is proposed. The algorithm relaxes the source independence assumption and has low-complexity algebraic com- putations. Simulation results on blind source separation including real face image data indicate that the sources can be successfully recovered with the algorithm.展开更多
In order to prevent possible casualties and economic loss, it is critical to accurate prediction of the Remaining Useful Life (RUL) in rail prognostics health management. However, the traditional neural networks is di...In order to prevent possible casualties and economic loss, it is critical to accurate prediction of the Remaining Useful Life (RUL) in rail prognostics health management. However, the traditional neural networks is difficult to capture the long-term dependency relationship of the time series in the modeling of the long time series of rail damage, due to the coupling relationship of multi-channel data from multiple sensors. Here, in this paper, a novel RUL prediction model with an enhanced pulse separable convolution is used to solve this issue. Firstly, a coding module based on the improved pulse separable convolutional network is established to effectively model the relationship between the data. To enhance the network, an alternate gradient back propagation method is implemented. And an efficient channel attention (ECA) mechanism is developed for better emphasizing the useful pulse characteristics. Secondly, an optimized Transformer encoder was designed to serve as the backbone of the model. It has the ability to efficiently understand relationship between the data itself and each other at each time step of long time series with a full life cycle. More importantly, the Transformer encoder is improved by integrating pulse maximum pooling to retain more pulse timing characteristics. Finally, based on the characteristics of the front layer, the final predicted RUL value was provided and served as the end-to-end solution. The empirical findings validate the efficacy of the suggested approach in forecasting the rail RUL, surpassing various existing data-driven prognostication techniques. Meanwhile, the proposed method also shows good generalization performance on PHM2012 bearing data set.展开更多
Deep kernel mapping support vector machines have achieved good results in numerous tasks by mapping features from a low-dimensional space to a high-dimensional space and then using support vector machines for classifi...Deep kernel mapping support vector machines have achieved good results in numerous tasks by mapping features from a low-dimensional space to a high-dimensional space and then using support vector machines for classification.However,the depth kernel mapping support vector machine does not take into account the connection of different dimensional spaces and increases the model parameters.To further improve the recognition capability of deep kernel mapping support vector machines while reducing the number of model parameters,this paper proposes a framework of Lightweight Deep Convolutional Cross-Connected Kernel Mapping Support Vector Machines(LC-CKMSVM).The framework consists of a feature extraction module and a classification module.The feature extraction module first maps the data from low-dimensional to high-dimensional space by fusing the representations of different dimensional spaces through cross-connections;then,it uses depthwise separable convolution to replace part of the original convolution to reduce the number of parameters in the module;The classification module uses a soft margin support vector machine for classification.The results on 6 different visual datasets show that LC-CKMSVM obtains better classification accuracies on most cases than the other five models.展开更多
One of the most obvious clinical reasons of dementia or The Behavioral and Psychological Symptoms of Dementia(BPSD)are the lack of emotional expression,the increased frequency of negative emotions,and the impermanence...One of the most obvious clinical reasons of dementia or The Behavioral and Psychological Symptoms of Dementia(BPSD)are the lack of emotional expression,the increased frequency of negative emotions,and the impermanence of emotions.Observing the reduction of BPSD in dementia through emotions can be considered effective and widely used in the field of non-pharmacological therapy.At present,this article will verify whether the image recognition artificial intelligence(AI)system can correctly reflect the emotional performance of the elderly with dementia through a questionnaire survey of three professional elderly nursing staff.The ANOVA(sig.=0.50)is used to determine that the judgment given by the nursing staff has no obvious deviation,and then Kendall's test(0.722**)and spearman's test(0.863**)are used to verify the judgment severity of the emotion recognition system and the nursing staff unanimously.This implies the usability of the tool.Additionally,it can be expected to be further applied in the research related to BPSD elderly emotion detection.展开更多
Multiple suppression is an important element of marine seismic data processing.Intelligent suppression of multiples us-ing artificial intelligence reduces labor costs,minimizes dependence on unknown prior information,...Multiple suppression is an important element of marine seismic data processing.Intelligent suppression of multiples us-ing artificial intelligence reduces labor costs,minimizes dependence on unknown prior information,and improves data processing ef-ficiency.In this study,we propose an intelligent method for suppressing marine seismic multiples using deep learning approaches.The proposed method enables the intelligent suppression of free-surface-related multiples from seismic records.Initially,we construct a multi-category marine seismic multiple dataset through finite difference forward modeling under different boundary conditions.We use various models and data augmentation methods,including sample rotation,noise addition,and random channel omission.Then,we apply depthwise separable convolution to develop our deep learning Mobilenet-Unet model.The Mobilenet-Unet framework sig-nificantly reduces the number of operations required for multiple elimination without sacrificing model performance,ultimately reali-zing the optimal multiple suppression model.The trained Mobilenet-Unet is applied to the test set for verification.Moreover,to deter-mine its generalization ability,it is implemented to seismic records containing multiples generated by two marine geophysical models that were not included in the training process.The performance of Mobilenet-Unet is also compared with that of different network structures.The results indicate that,despite its small size,our proposed Mobilenet-Unet deep learning model can rapidly and effective-ly separate multiples in marine seismic data,possessing reasonable generalization ability.展开更多
To address the limitations of existing abnormal traffic detection methods,such as insufficient temporal and spatial feature extraction,high false positive rate(FPR),poor generalization,and class imbalance,this study p...To address the limitations of existing abnormal traffic detection methods,such as insufficient temporal and spatial feature extraction,high false positive rate(FPR),poor generalization,and class imbalance,this study proposed an intelligent detection method that combines a Stacked Convolutional Network(SCN),Bidirectional Long Short-Term Memory(BiLSTM)network,and Equalization Loss v2(EQL v2).This method was divided into two components:a feature extraction model and a classification and detection model.First,SCN was constructed by combining a Convolutional Neural Network(CNN)with a Depthwise Separable Convolution(DSC)network to capture the abstract spatial features of traffic data.These features were then input into the BiLSTM to capture temporal dependencies.An attention mechanism was incorporated after SCN and BiLSTM to enhance the extraction of key spatiotemporal features.To address class imbalance,the classification detection model applied EQL v2 to adjust the weights of the minority classes,ensuring that they received equal focus during training.The experimental results indicated that the proposed method outperformed the existing methods in terms of accuracy,FPR,and F1-score and significantly improved the identification rate of minority classes.展开更多
The accurate and automatic segmentation of retinal vessels fromfundus images is critical for the early diagnosis and prevention ofmany eye diseases,such as diabetic retinopathy(DR).Existing retinal vessel segmentation...The accurate and automatic segmentation of retinal vessels fromfundus images is critical for the early diagnosis and prevention ofmany eye diseases,such as diabetic retinopathy(DR).Existing retinal vessel segmentation approaches based on convolutional neural networks(CNNs)have achieved remarkable effectiveness.Here,we extend a retinal vessel segmentation model with low complexity and high performance based on U-Net,which is one of the most popular architectures.In view of the excellent work of depth-wise separable convolution,we introduce it to replace the standard convolutional layer.The complexity of the proposed model is reduced by decreasing the number of parameters and calculations required for themodel.To ensure performance while lowering redundant parameters,we integrate the pre-trained MobileNet V2 into the encoder.Then,a feature fusion residual module(FFRM)is designed to facilitate complementary strengths by enhancing the effective fusion between adjacent levels,which alleviates extraneous clutter introduced by direct fusion.Finally,we provide detailed comparisons between the proposed SepFE and U-Net in three retinal image mainstream datasets(DRIVE,STARE,and CHASEDB1).The results show that the number of SepFE parameters is only 3%of U-Net,the Flops are only 8%of U-Net,and better segmentation performance is obtained.The superiority of SepFE is further demonstrated through comparisons with other advanced methods.展开更多
Purpose:As to January 11,2021,coronavirus disease(COVID-19)has caused more than 2 million deaths worldwide.Mainly diagnostic methods of COVID-19 are:(i)nucleic acid testing.This method requires high requirements on th...Purpose:As to January 11,2021,coronavirus disease(COVID-19)has caused more than 2 million deaths worldwide.Mainly diagnostic methods of COVID-19 are:(i)nucleic acid testing.This method requires high requirements on the sample testing environment.When collecting samples,staff are in a susceptible environment,which increases the risk of infection.(ii)chest computed tomography.The cost of it is high and some radiation in the scan process.(iii)chest X-ray images.It has the advantages of fast imaging,higher spatial recognition than chest computed tomography.Therefore,our team chose the chest X-ray images as the experimental dataset in this paper.Methods:We proposed a novel framework—BEVGG and three methods(BEVGGC-I,BEVGGC-II,and BEVGGC-III)to diagnose COVID-19 via chest X-ray images.Besides,we used biogeography-based optimization to optimize the values of hyperparameters of the convolutional neural network.Results:The experimental results show that the OA of our proposed three methods are 97.65%±0.65%,94.49%±0.22%and 94.81%±0.52%.BEVGGC-I has the best performance of all methods.Conclusions:The OA of BEVGGC-I is 9.59%±1.04%higher than that of state-of-the-art methods.展开更多
Image semantic segmentation has become an essential part of autonomous driving.To further improve the generalization ability and the robustness of semantic segmentation algorithms,a lightweight algorithm network based...Image semantic segmentation has become an essential part of autonomous driving.To further improve the generalization ability and the robustness of semantic segmentation algorithms,a lightweight algorithm network based on Squeeze-and-Excitation Attention Mechanism(SE)and Depthwise Separable Convolution(DSC)is designed.Meanwhile,Adam-GC,an Adam optimization algorithm based on Gradient Compression(GC),is proposed to improve the training speed,segmentation accuracy,generalization ability and stability of the algorithm network.To verify and compare the effectiveness of the algorithm network proposed in this paper,the trained networkmodel is used for experimental verification and comparative test on the Cityscapes semantic segmentation dataset.The validation and comparison results show that the overall segmentation results of the algorithmnetwork can achieve 78.02%MIoU on Cityscapes validation set,which is better than the basic algorithm network and the other latest semantic segmentation algorithms network.Besides meeting the stability and accuracy requirements,it has a particular significance for the development of image semantic segmentation.展开更多
In response to the problem of the high cost and low efficiency of traditional water surface litter cleanup through manpower,a lightweight water surface litter detection algorithm based on improved YOLOv5s is proposed ...In response to the problem of the high cost and low efficiency of traditional water surface litter cleanup through manpower,a lightweight water surface litter detection algorithm based on improved YOLOv5s is proposed to provide core technical support for real-time water surface litter detection by water surface litter cleanup vessels.The method reduces network parameters by introducing the deep separable convolution GhostConv in the lightweight network GhostNet to substitute the ordinary convolution in the original YOLOv5s feature extraction and fusion network;introducing the C3Ghost module to substitute the C3 module in the original backbone and neck networks to further reduce computational effort.Using a Convolutional Block Attention Mechanism(CBAM)module in the backbone network to strengthen the network’s ability to extract significant target features from images.Finally,the loss function is optimized using the Focal-EIoU loss func-tion to improve the convergence speed and model accuracy.The experimental results illustrate that the improved algorithm outperforms the original Yolov5s in all aspects of the homemade water surface litter dataset and has certain advantages over some current mainstream algorithms in terms of model size,detection accuracy,and speed,which can deal with the problems of real-time detection of water surface litter in real life.展开更多
Infrared target detection models are more required than ever before to be deployed on embedded platforms,which requires models with less memory consumption and better real-time performance while considering accuracy.T...Infrared target detection models are more required than ever before to be deployed on embedded platforms,which requires models with less memory consumption and better real-time performance while considering accuracy.To address the above challenges,we propose a modified You Only Look Once(YOLO)algorithm PF-YOLOv4-Tiny.The algorithm incorpo-rates spatial pyramidal pooling(SPP)and squeeze-and-excitation(SE)visual attention modules to enhance the target localization capability.The PANet-based-feature pyramid networks(P-FPN)are proposed to transfer semantic information and location information simultaneously to ameliorate detection accuracy.To lighten the network,the standard convolutions other than the backbone network are replaced with depthwise separable convolutions.In post-processing the images,the soft-non-maximum suppression(soft-NMS)algorithm is employed to subside the missed and false detection problems caused by the occlusion between targets.The accuracy of our model can finally reach 61.75%,while the total Params is only 9.3 M and GFLOPs is 11.At the same time,the inference speed reaches 87 FPS on NVIDIA GeForce GTX 1650 Ti,which can meet the requirements of the infrared target detection algorithm for the embedded deployments.展开更多
With the rapid development of social economy,transportation has become faster and more efficient.As an important part of goods transportation,the safe maintenance of tunnel highways has become particularly important.T...With the rapid development of social economy,transportation has become faster and more efficient.As an important part of goods transportation,the safe maintenance of tunnel highways has become particularly important.The maintenance of tunnel roads has become more difficult due to problems such as sealing,narrowness and lack of light.Currently,target detection methods are advantageous in detecting tunnel vehicles in a timely manner through monitoring.Therefore,in order to prevent vehicle misdetection and missed detection in this complex environment,we propose aYOLOv5-Vehicle model based on the YOLOv5 network.This model is improved in three ways.Firstly,The backbone network of YOLOv5 is replaced by the lightweight MobileNetV3 network to extract features,which reduces the number of model parameters;Next,all convolutions in the neck module are improved to the depth-wise separable convolutions to further reduce the number of model parameters and computation,and improve the detection speed of the model;Finally,to ensure the accuracy of the model,the CBAM attention mechanism is introduced to improve the detection accuracy and precision of the model.Experiments results demonstrate that the YOLOv5-Vehicle model can improve the accuracy.展开更多
According to recent research statistics,approximately 30%of people who experienced falls are over the age of 65.Therefore,it is meaningful research to detect it in time and take appropriate measures when falling behav...According to recent research statistics,approximately 30%of people who experienced falls are over the age of 65.Therefore,it is meaningful research to detect it in time and take appropriate measures when falling behavior occurs.In this paper,a fall detection model based on improved human posture estimation algorithm is proposed.The improved human posture estimation algorithm is implemented on the basis of Openpose.An im-proved strategy based on depthwise separable convolution combined with HDC structure is proposed.The depthwise separable convolution is used to replace the convolution neural network structure,which makes the network lightweight and reduces the redundant layer in the network.At the same time,in order to ensure that the image features are not lost and ensure the accuracy of detecting human joint points,HDC structure is introduced.Experiments show that the improved algorithm with HDC structure has higher accuracy in joint point detection.Then,human posture estimation is applied to fall detection research,and fall event modeling is carried out through fall feature extraction.The designed convolution neural network model is used to classify and distinguish falls.The experimental results show that our method achieves 98.53%,97.71%and 97.20%accuracy on three public fall detection data sets.Compared with the experimental results of other methods on the same data set,the model designed in this paper has a certain improvement in system accuracy.The sensitivity is also improved,which will reduce the error detection probability of the system.In addition,this paper also verifies the real-time performance of the model.Even if researchers are experimenting with low-level hardware,it can ensure a certain detection speed without too much delay.展开更多
Memristor-based neuromorphic computing shows great potential for high-speed and high-throughput signal processing applications,such as electroencephalogram(EEG)signal processing.Nonetheless,the size of one-transistor ...Memristor-based neuromorphic computing shows great potential for high-speed and high-throughput signal processing applications,such as electroencephalogram(EEG)signal processing.Nonetheless,the size of one-transistor one-resistor(1T1R)memristor arrays is limited by the non-ideality of the devices,which prevents the hardware implementation of large and complex networks.In this work,we propose the depthwise separable convolution and bidirectional gate recurrent unit(DSC-BiGRU)network,a lightweight and highly robust hybrid neural network based on 1T1R arrays that enables efficient processing of EEG signals in the temporal,frequency and spatial domains by hybridizing DSC and BiGRU blocks.The network size is reduced and the network robustness is improved while ensuring the network classification accuracy.In the simulation,the measured non-idealities of the 1T1R array are brought into the network through statistical analysis.Compared with traditional convolutional networks,the network parameters are reduced by 95%and the network classification accuracy is improved by 21%at a 95%array yield rate and 5%tolerable error.This work demonstrates that lightweight and highly robust networks based on memristor arrays hold great promise for applications that rely on low consumption and high efficiency.展开更多
In order to solve the problems of shallow features loss and high computation cost of U-Net,we propose a lightweight with shallow features combination(IU-Net).IU-Net adds several convolution layers and short links to t...In order to solve the problems of shallow features loss and high computation cost of U-Net,we propose a lightweight with shallow features combination(IU-Net).IU-Net adds several convolution layers and short links to the skip path to extract more shallow features.At the same time,the original convolution is replaced by the depth-wise separable convolution to reduce the calculation cost and the number of parameters.IU-Net is applied to detecting small metal industrial products defects.It is evaluated on our own SUES-Washer dataset to verify the effectiveness.Experimental results demonstrate that our proposed method outperforms the original U-Net,and it has 1.73%,2.08%and 11.2%improvement in the intersection over union,accuracy,and detection time,respectively,which satisfies the requirements of industrial detection.展开更多
基金supported by the National Natural Science Foundation of China(No.52277055).
文摘Traditional data-driven fault diagnosis methods depend on expert experience to manually extract effective fault features of signals,which has certain limitations.Conversely,deep learning techniques have gained prominence as a central focus of research in the field of fault diagnosis by strong fault feature extraction ability and end-to-end fault diagnosis efficiency.Recently,utilizing the respective advantages of convolution neural network(CNN)and Transformer in local and global feature extraction,research on cooperating the two have demonstrated promise in the field of fault diagnosis.However,the cross-channel convolution mechanism in CNN and the self-attention calculations in Transformer contribute to excessive complexity in the cooperative model.This complexity results in high computational costs and limited industrial applicability.To tackle the above challenges,this paper proposes a lightweight CNN-Transformer named as SEFormer for rotating machinery fault diagnosis.First,a separable multiscale depthwise convolution block is designed to extract and integrate multiscale feature information from different channel dimensions of vibration signals.Then,an efficient self-attention block is developed to capture critical fine-grained features of the signal from a global perspective.Finally,experimental results on the planetary gearbox dataset and themotor roller bearing dataset prove that the proposed framework can balance the advantages of robustness,generalization and lightweight compared to recent state-of-the-art fault diagnosis models based on CNN and Transformer.This study presents a feasible strategy for developing a lightweight rotating machinery fault diagnosis framework aimed at economical deployment.
基金This work was supported by Liaoning Provincial Science Public Welfare Research Fund Project(No.2016002006)Liaoning Provincial Department of Education Scientific Research Service Local Project(No.L201708).
文摘Recently,video-based fire detection technology has become an important research topic in the field of machine vision.This paper proposes a method of combining the classification model and target detection model in deep learning for fire detection.Firstly,the depthwise separable convolution is used to classify fire images,which saves a lot of detection time under the premise of ensuring detection accuracy.Secondly,You Only Look Once version 3(YOLOv3)target regression function is used to output the fire position information for the images whose classification result is fire,which avoids the problem that the accuracy of detection cannot be guaranteed by using YOLOv3 for target classification and position regression.At the same time,the detection time of target regression for images without fire is greatly reduced saved.The experiments were tested using a network public database.The detection accuracy reached 98%and the detection rate reached 38fps.This method not only saves the workload of manually extracting flame characteristics,reduces the calculation cost,and reduces the amount of parameters,but also improves the detection accuracy and detection rate.
基金the grant from the Ph.D. Programs Foun-dation of Ministry of Education of China (No. 20060280003)the Shanghai Leading Academic Dis-cipline Project (Project No.T0102).
文摘This letter deals with the frequency domain Blind Source Separation of Convolutive Mixtures (CMBSS). From the frequency representation of the "overlap and save", a Weighted General Discrete Fourier Transform (WGDFT) is derived to replace the traditional Discrete Fourier Transform (DFT). The mixing matrix on each frequency bin could be estimated more precisely from WGDFT coefficients than from DFT coefficients, which improves separation performance. Simulation results verify the validity of WGDFT for frequency domain blind source separation of convolutive mixtures.
文摘Convolutional neural networks (CNNs) are widely used in image classification tasks, but their increasing model size and computation make them challenging to implement on embedded systems with constrained hardware resources. To address this issue, the MobileNetV1 network was developed, which employs depthwise convolution to reduce network complexity. MobileNetV1 employs a stride of 2 in several convolutional layers to decrease the spatial resolution of feature maps, thereby lowering computational costs. However, this stride setting can lead to a loss of spatial information, particularly affecting the detection and representation of smaller objects or finer details in images. To maintain the trade-off between complexity and model performance, a lightweight convolutional neural network with hierarchical multi-scale feature fusion based on the MobileNetV1 network is proposed. The network consists of two main subnetworks. The first subnetwork uses a depthwise dilated separable convolution (DDSC) layer to learn imaging features with fewer parameters, which results in a lightweight and computationally inexpensive network. Furthermore, depthwise dilated convolution in DDSC layer effectively expands the field of view of filters, allowing them to incorporate a larger context. The second subnetwork is a hierarchical multi-scale feature fusion (HMFF) module that uses parallel multi-resolution branches architecture to process the input feature map in order to extract the multi-scale feature information of the input image. Experimental results on the CIFAR-10, Malaria, and KvasirV1 datasets demonstrate that the proposed method is efficient, reducing the network parameters and computational cost by 65.02% and 39.78%, respectively, while maintaining the network performance compared to the MobileNetV1 baseline.
基金Supported by the National Natural Science Foundation of China(61903336,61976190)the Natural Science Foundation of Zhejiang Province(LY21F030015)。
文摘Background The use of remote photoplethysmography(rPPG)to estimate blood volume pulse in a noncontact manner has been an active research topic in recent years.Existing methods are primarily based on a singlescale region of interest(ROI).However,some noise signals that are not easily separated in a single-scale space can be easily separated in a multi-scale space.Also,existing spatiotemporal networks mainly focus on local spatiotemporal information and do not emphasize temporal information,which is crucial in pulse extraction problems,resulting in insufficient spatiotemporal feature modelling.Methods Here,we propose a multi-scale facial video pulse extraction network based on separable spatiotemporal convolution(SSTC)and dimension separable attention(DSAT).First,to solve the problem of a single-scale ROI,we constructed a multi-scale feature space for initial signal separation.Second,SSTC and DSAT were designed for efficient spatiotemporal correlation modeling,which increased the information interaction between the long-span time and space dimensions;this placed more emphasis on temporal features.Results The signal-to-noise ratio(SNR)of the proposed network reached 9.58dB on the PURE dataset and 6.77dB on the UBFC-rPPG dataset,outperforming state-of-the-art algorithms.Conclusions The results showed that fusing multi-scale signals yielded better results than methods based on only single-scale signals.The proposed SSTC and dimension-separable attention mechanism will contribute to more accurate pulse signal extraction.
基金Supported by the Specialized Research Fund for the Doctoral Program of Higher Education of China (No.20060280003)Shanghai Leading Academic Dis-cipline Project (T0102)
文摘Most of the existing algorithms for blind sources separation have a limitation that sources are statistically independent. However, in many practical applications, the source signals are non- negative and mutual statistically dependent signals. When the observations are nonnegative linear combinations of nonnegative sources, the correlation coefficients of the observations are larger than these of source signals. In this letter, a novel Nonnegative Matrix Factorization (NMF) algorithm with least correlated component constraints to blind separation of convolutive mixed sources is proposed. The algorithm relaxes the source independence assumption and has low-complexity algebraic com- putations. Simulation results on blind source separation including real face image data indicate that the sources can be successfully recovered with the algorithm.
文摘In order to prevent possible casualties and economic loss, it is critical to accurate prediction of the Remaining Useful Life (RUL) in rail prognostics health management. However, the traditional neural networks is difficult to capture the long-term dependency relationship of the time series in the modeling of the long time series of rail damage, due to the coupling relationship of multi-channel data from multiple sensors. Here, in this paper, a novel RUL prediction model with an enhanced pulse separable convolution is used to solve this issue. Firstly, a coding module based on the improved pulse separable convolutional network is established to effectively model the relationship between the data. To enhance the network, an alternate gradient back propagation method is implemented. And an efficient channel attention (ECA) mechanism is developed for better emphasizing the useful pulse characteristics. Secondly, an optimized Transformer encoder was designed to serve as the backbone of the model. It has the ability to efficiently understand relationship between the data itself and each other at each time step of long time series with a full life cycle. More importantly, the Transformer encoder is improved by integrating pulse maximum pooling to retain more pulse timing characteristics. Finally, based on the characteristics of the front layer, the final predicted RUL value was provided and served as the end-to-end solution. The empirical findings validate the efficacy of the suggested approach in forecasting the rail RUL, surpassing various existing data-driven prognostication techniques. Meanwhile, the proposed method also shows good generalization performance on PHM2012 bearing data set.
基金This work is supported by the National Natural Science Foundation of China(61806013,61876010,61906005,62166002)General project of Science and Technology Plan of Beijing Municipal Education Commission(KM202110005028)+1 种基金Project of Interdisciplinary Research Institute of Beijing University of Technology(2021020101)International Research Cooperation Seed Fund of Beijing University of Technology(2021A01).
文摘Deep kernel mapping support vector machines have achieved good results in numerous tasks by mapping features from a low-dimensional space to a high-dimensional space and then using support vector machines for classification.However,the depth kernel mapping support vector machine does not take into account the connection of different dimensional spaces and increases the model parameters.To further improve the recognition capability of deep kernel mapping support vector machines while reducing the number of model parameters,this paper proposes a framework of Lightweight Deep Convolutional Cross-Connected Kernel Mapping Support Vector Machines(LC-CKMSVM).The framework consists of a feature extraction module and a classification module.The feature extraction module first maps the data from low-dimensional to high-dimensional space by fusing the representations of different dimensional spaces through cross-connections;then,it uses depthwise separable convolution to replace part of the original convolution to reduce the number of parameters in the module;The classification module uses a soft margin support vector machine for classification.The results on 6 different visual datasets show that LC-CKMSVM obtains better classification accuracies on most cases than the other five models.
文摘One of the most obvious clinical reasons of dementia or The Behavioral and Psychological Symptoms of Dementia(BPSD)are the lack of emotional expression,the increased frequency of negative emotions,and the impermanence of emotions.Observing the reduction of BPSD in dementia through emotions can be considered effective and widely used in the field of non-pharmacological therapy.At present,this article will verify whether the image recognition artificial intelligence(AI)system can correctly reflect the emotional performance of the elderly with dementia through a questionnaire survey of three professional elderly nursing staff.The ANOVA(sig.=0.50)is used to determine that the judgment given by the nursing staff has no obvious deviation,and then Kendall's test(0.722**)and spearman's test(0.863**)are used to verify the judgment severity of the emotion recognition system and the nursing staff unanimously.This implies the usability of the tool.Additionally,it can be expected to be further applied in the research related to BPSD elderly emotion detection.
基金supported by the Key Laboratory of Ma-rine Mineral Resources,Ministry of Natural Resources,Guangzhou(No.KLMMR-2022-G09)the Guangzhou Ba-sic Research Program-Basic and Basic Applied Research Project(No.2023A04J0917)the PI Project of South-ern Marine Science and Engineering Guangdong Labora-tory(Guangzhou)(No.GML2020GD0802).
文摘Multiple suppression is an important element of marine seismic data processing.Intelligent suppression of multiples us-ing artificial intelligence reduces labor costs,minimizes dependence on unknown prior information,and improves data processing ef-ficiency.In this study,we propose an intelligent method for suppressing marine seismic multiples using deep learning approaches.The proposed method enables the intelligent suppression of free-surface-related multiples from seismic records.Initially,we construct a multi-category marine seismic multiple dataset through finite difference forward modeling under different boundary conditions.We use various models and data augmentation methods,including sample rotation,noise addition,and random channel omission.Then,we apply depthwise separable convolution to develop our deep learning Mobilenet-Unet model.The Mobilenet-Unet framework sig-nificantly reduces the number of operations required for multiple elimination without sacrificing model performance,ultimately reali-zing the optimal multiple suppression model.The trained Mobilenet-Unet is applied to the test set for verification.Moreover,to deter-mine its generalization ability,it is implemented to seismic records containing multiples generated by two marine geophysical models that were not included in the training process.The performance of Mobilenet-Unet is also compared with that of different network structures.The results indicate that,despite its small size,our proposed Mobilenet-Unet deep learning model can rapidly and effective-ly separate multiples in marine seismic data,possessing reasonable generalization ability.
基金supported by the National Natural Science Foundation of China(Grant No.62102449).
文摘To address the limitations of existing abnormal traffic detection methods,such as insufficient temporal and spatial feature extraction,high false positive rate(FPR),poor generalization,and class imbalance,this study proposed an intelligent detection method that combines a Stacked Convolutional Network(SCN),Bidirectional Long Short-Term Memory(BiLSTM)network,and Equalization Loss v2(EQL v2).This method was divided into two components:a feature extraction model and a classification and detection model.First,SCN was constructed by combining a Convolutional Neural Network(CNN)with a Depthwise Separable Convolution(DSC)network to capture the abstract spatial features of traffic data.These features were then input into the BiLSTM to capture temporal dependencies.An attention mechanism was incorporated after SCN and BiLSTM to enhance the extraction of key spatiotemporal features.To address class imbalance,the classification detection model applied EQL v2 to adjust the weights of the minority classes,ensuring that they received equal focus during training.The experimental results indicated that the proposed method outperformed the existing methods in terms of accuracy,FPR,and F1-score and significantly improved the identification rate of minority classes.
基金supported by the Hunan Provincial Natural Science Foundation of China(2021JJ50074)the Scientific Research Fund of Hunan Provincial Education Department(19B082)+6 种基金the Science and Technology Development Center of the Ministry of Education-New Generation Information Technology Innovation Project(2018A02020)the Science Foundation of Hengyang Normal University(19QD12)the Science and Technology Plan Project of Hunan Province(2016TP1020)the Subject Group Construction Project of Hengyang Normal University(18XKQ02)theApplication Oriented SpecialDisciplines,Double First ClassUniversity Project of Hunan Province(Xiangjiaotong[2018]469)the Hunan Province Special Funds of Central Government for Guiding Local Science and Technology Development(2018CT5001)the First Class Undergraduate Major in Hunan Province Internet of Things Major(Xiangjiaotong[2020]248,No.288).
文摘The accurate and automatic segmentation of retinal vessels fromfundus images is critical for the early diagnosis and prevention ofmany eye diseases,such as diabetic retinopathy(DR).Existing retinal vessel segmentation approaches based on convolutional neural networks(CNNs)have achieved remarkable effectiveness.Here,we extend a retinal vessel segmentation model with low complexity and high performance based on U-Net,which is one of the most popular architectures.In view of the excellent work of depth-wise separable convolution,we introduce it to replace the standard convolutional layer.The complexity of the proposed model is reduced by decreasing the number of parameters and calculations required for themodel.To ensure performance while lowering redundant parameters,we integrate the pre-trained MobileNet V2 into the encoder.Then,a feature fusion residual module(FFRM)is designed to facilitate complementary strengths by enhancing the effective fusion between adjacent levels,which alleviates extraneous clutter introduced by direct fusion.Finally,we provide detailed comparisons between the proposed SepFE and U-Net in three retinal image mainstream datasets(DRIVE,STARE,and CHASEDB1).The results show that the number of SepFE parameters is only 3%of U-Net,the Flops are only 8%of U-Net,and better segmentation performance is obtained.The superiority of SepFE is further demonstrated through comparisons with other advanced methods.
基金Key Science and Technology Program of Henan Province,China(212102310084)J.Sun,X.Li,and C.Tang received the grant.Provincial Key Laboratory for Computer Information Processing Technology,Soochow University(KJS2048),J.Sun received the grant.
文摘Purpose:As to January 11,2021,coronavirus disease(COVID-19)has caused more than 2 million deaths worldwide.Mainly diagnostic methods of COVID-19 are:(i)nucleic acid testing.This method requires high requirements on the sample testing environment.When collecting samples,staff are in a susceptible environment,which increases the risk of infection.(ii)chest computed tomography.The cost of it is high and some radiation in the scan process.(iii)chest X-ray images.It has the advantages of fast imaging,higher spatial recognition than chest computed tomography.Therefore,our team chose the chest X-ray images as the experimental dataset in this paper.Methods:We proposed a novel framework—BEVGG and three methods(BEVGGC-I,BEVGGC-II,and BEVGGC-III)to diagnose COVID-19 via chest X-ray images.Besides,we used biogeography-based optimization to optimize the values of hyperparameters of the convolutional neural network.Results:The experimental results show that the OA of our proposed three methods are 97.65%±0.65%,94.49%±0.22%and 94.81%±0.52%.BEVGGC-I has the best performance of all methods.Conclusions:The OA of BEVGGC-I is 9.59%±1.04%higher than that of state-of-the-art methods.
基金supported by Qingdao People’s Livelihood Science and Technology Plan (Grant 19-6-1-88-nsh).
文摘Image semantic segmentation has become an essential part of autonomous driving.To further improve the generalization ability and the robustness of semantic segmentation algorithms,a lightweight algorithm network based on Squeeze-and-Excitation Attention Mechanism(SE)and Depthwise Separable Convolution(DSC)is designed.Meanwhile,Adam-GC,an Adam optimization algorithm based on Gradient Compression(GC),is proposed to improve the training speed,segmentation accuracy,generalization ability and stability of the algorithm network.To verify and compare the effectiveness of the algorithm network proposed in this paper,the trained networkmodel is used for experimental verification and comparative test on the Cityscapes semantic segmentation dataset.The validation and comparison results show that the overall segmentation results of the algorithmnetwork can achieve 78.02%MIoU on Cityscapes validation set,which is better than the basic algorithm network and the other latest semantic segmentation algorithms network.Besides meeting the stability and accuracy requirements,it has a particular significance for the development of image semantic segmentation.
基金Support for this work was in part from the China University Industry-University Research Innovation Fund Project(No.2022BL052),author B.T,https://www.cutech.edu.cnin part by the Science and Technology InnovationR&DProject of the State GeneralAdministration of Sports of China(No.22KJCX024),author B.T,https://www.sport.gov.cn+1 种基金in part by the Major Project of Philosophy and Social Science Research in Higher Education Institutions in Hubei Province(No.21ZD054),author B.T,https://jyt.hubei.gov.cnKey Project of Hubei Provincial Key Laboratory of Intelligent Transportation Technology and Equipment Open Fund(No.2022XZ106),author B.T,https://hbpu.edu.cn.
文摘In response to the problem of the high cost and low efficiency of traditional water surface litter cleanup through manpower,a lightweight water surface litter detection algorithm based on improved YOLOv5s is proposed to provide core technical support for real-time water surface litter detection by water surface litter cleanup vessels.The method reduces network parameters by introducing the deep separable convolution GhostConv in the lightweight network GhostNet to substitute the ordinary convolution in the original YOLOv5s feature extraction and fusion network;introducing the C3Ghost module to substitute the C3 module in the original backbone and neck networks to further reduce computational effort.Using a Convolutional Block Attention Mechanism(CBAM)module in the backbone network to strengthen the network’s ability to extract significant target features from images.Finally,the loss function is optimized using the Focal-EIoU loss func-tion to improve the convergence speed and model accuracy.The experimental results illustrate that the improved algorithm outperforms the original Yolov5s in all aspects of the homemade water surface litter dataset and has certain advantages over some current mainstream algorithms in terms of model size,detection accuracy,and speed,which can deal with the problems of real-time detection of water surface litter in real life.
基金supported by The Natural Science Foundation of the Jiangsu Higher Education Institutions of China(Grants No.19JKB520031).
文摘Infrared target detection models are more required than ever before to be deployed on embedded platforms,which requires models with less memory consumption and better real-time performance while considering accuracy.To address the above challenges,we propose a modified You Only Look Once(YOLO)algorithm PF-YOLOv4-Tiny.The algorithm incorpo-rates spatial pyramidal pooling(SPP)and squeeze-and-excitation(SE)visual attention modules to enhance the target localization capability.The PANet-based-feature pyramid networks(P-FPN)are proposed to transfer semantic information and location information simultaneously to ameliorate detection accuracy.To lighten the network,the standard convolutions other than the backbone network are replaced with depthwise separable convolutions.In post-processing the images,the soft-non-maximum suppression(soft-NMS)algorithm is employed to subside the missed and false detection problems caused by the occlusion between targets.The accuracy of our model can finally reach 61.75%,while the total Params is only 9.3 M and GFLOPs is 11.At the same time,the inference speed reaches 87 FPS on NVIDIA GeForce GTX 1650 Ti,which can meet the requirements of the infrared target detection algorithm for the embedded deployments.
文摘With the rapid development of social economy,transportation has become faster and more efficient.As an important part of goods transportation,the safe maintenance of tunnel highways has become particularly important.The maintenance of tunnel roads has become more difficult due to problems such as sealing,narrowness and lack of light.Currently,target detection methods are advantageous in detecting tunnel vehicles in a timely manner through monitoring.Therefore,in order to prevent vehicle misdetection and missed detection in this complex environment,we propose aYOLOv5-Vehicle model based on the YOLOv5 network.This model is improved in three ways.Firstly,The backbone network of YOLOv5 is replaced by the lightweight MobileNetV3 network to extract features,which reduces the number of model parameters;Next,all convolutions in the neck module are improved to the depth-wise separable convolutions to further reduce the number of model parameters and computation,and improve the detection speed of the model;Finally,to ensure the accuracy of the model,the CBAM attention mechanism is introduced to improve the detection accuracy and precision of the model.Experiments results demonstrate that the YOLOv5-Vehicle model can improve the accuracy.
文摘According to recent research statistics,approximately 30%of people who experienced falls are over the age of 65.Therefore,it is meaningful research to detect it in time and take appropriate measures when falling behavior occurs.In this paper,a fall detection model based on improved human posture estimation algorithm is proposed.The improved human posture estimation algorithm is implemented on the basis of Openpose.An im-proved strategy based on depthwise separable convolution combined with HDC structure is proposed.The depthwise separable convolution is used to replace the convolution neural network structure,which makes the network lightweight and reduces the redundant layer in the network.At the same time,in order to ensure that the image features are not lost and ensure the accuracy of detecting human joint points,HDC structure is introduced.Experiments show that the improved algorithm with HDC structure has higher accuracy in joint point detection.Then,human posture estimation is applied to fall detection research,and fall event modeling is carried out through fall feature extraction.The designed convolution neural network model is used to classify and distinguish falls.The experimental results show that our method achieves 98.53%,97.71%and 97.20%accuracy on three public fall detection data sets.Compared with the experimental results of other methods on the same data set,the model designed in this paper has a certain improvement in system accuracy.The sensitivity is also improved,which will reduce the error detection probability of the system.In addition,this paper also verifies the real-time performance of the model.Even if researchers are experimenting with low-level hardware,it can ensure a certain detection speed without too much delay.
基金Project supported by the National Key Research and Development Program of China(Grant No.2019YFB2205102)the National Natural Science Foundation of China(Grant Nos.61974164,62074166,61804181,62004219,62004220,and 62104256).
文摘Memristor-based neuromorphic computing shows great potential for high-speed and high-throughput signal processing applications,such as electroencephalogram(EEG)signal processing.Nonetheless,the size of one-transistor one-resistor(1T1R)memristor arrays is limited by the non-ideality of the devices,which prevents the hardware implementation of large and complex networks.In this work,we propose the depthwise separable convolution and bidirectional gate recurrent unit(DSC-BiGRU)network,a lightweight and highly robust hybrid neural network based on 1T1R arrays that enables efficient processing of EEG signals in the temporal,frequency and spatial domains by hybridizing DSC and BiGRU blocks.The network size is reduced and the network robustness is improved while ensuring the network classification accuracy.In the simulation,the measured non-idealities of the 1T1R array are brought into the network through statistical analysis.Compared with traditional convolutional networks,the network parameters are reduced by 95%and the network classification accuracy is improved by 21%at a 95%array yield rate and 5%tolerable error.This work demonstrates that lightweight and highly robust networks based on memristor arrays hold great promise for applications that rely on low consumption and high efficiency.
基金the Youth Fund of National Natural Science Foundation of China(61801286,62006150)Shanghai Young Science and Technology Talents Sailing Program(19YF1418400)Fund Project of Shanghai Science and Technology Commission(16dz1206002)。
文摘In order to solve the problems of shallow features loss and high computation cost of U-Net,we propose a lightweight with shallow features combination(IU-Net).IU-Net adds several convolution layers and short links to the skip path to extract more shallow features.At the same time,the original convolution is replaced by the depth-wise separable convolution to reduce the calculation cost and the number of parameters.IU-Net is applied to detecting small metal industrial products defects.It is evaluated on our own SUES-Washer dataset to verify the effectiveness.Experimental results demonstrate that our proposed method outperforms the original U-Net,and it has 1.73%,2.08%and 11.2%improvement in the intersection over union,accuracy,and detection time,respectively,which satisfies the requirements of industrial detection.
