Human disturbance activities is one of the main reasons for inducing geohazards.Ecological impact assessment metrics of roads are inconsistent criteria and multiple.From the perspective of visual observation,the envir...Human disturbance activities is one of the main reasons for inducing geohazards.Ecological impact assessment metrics of roads are inconsistent criteria and multiple.From the perspective of visual observation,the environment damage can be shown through detecting the uncovered area of vegetation in the images along road.To realize this,an end-to-end environment damage detection model based on convolutional neural network is proposed.A 50-layer residual network is used to extract feature map.The initial parameters are optimized by transfer learning.An example is shown by this method.The dataset including cliff and landslide damage are collected by us along road in Shennongjia national forest park.Results show 0.4703 average precision(AP)rating for cliff damage and 0.4809 average precision(AP)rating for landslide damage.Compared with YOLOv3,our model shows a better accuracy in cliff and landslide detection although a certain amount of speed is sacrificed.展开更多
Non-orthogonal multiple access(NOMA), featuring high spectrum efficiency, massive connectivity and low latency, holds immense potential to be a novel multi-access technique in fifth-generation(5G) communication. Succe...Non-orthogonal multiple access(NOMA), featuring high spectrum efficiency, massive connectivity and low latency, holds immense potential to be a novel multi-access technique in fifth-generation(5G) communication. Successive interference cancellation(SIC) is proved to be an effective method to detect the NOMA signal by ordering the power of received signals and then decoding them. However, the error accumulation effect referred to as error propagation is an inevitable problem. In this paper,we propose a convolutional neural networks(CNNs) approach to restore the desired signal impaired by the multiple input multiple output(MIMO) channel. Especially in the uplink NOMA scenario,the proposed method can decode multiple users' information in a cluster instantaneously without any traditional communication signal processing steps. Simulation experiments are conducted in the Rayleigh channel and the results demonstrate that the error performance of the proposed learning system outperforms that of the classic SIC detection. Consequently, deep learning has disruptive potential to replace the conventional signal detection method.展开更多
High-resolution flow field data has important applications in meteorology,aerospace engineering,high-energy physics and other fields.Experiments and numerical simulations are two main ways to obtain high-resolution fl...High-resolution flow field data has important applications in meteorology,aerospace engineering,high-energy physics and other fields.Experiments and numerical simulations are two main ways to obtain high-resolution flow field data,while the high experiment cost and computing resources for simulation hinder the specificanalysis of flow field evolution.With the development of deep learning technology,convolutional neural networks areused to achieve high-resolution reconstruction of the flow field.In this paper,an ordinary convolutional neuralnetwork and a multi-time-path convolutional neural network are established for the ablative Rayleigh-Taylorinstability.These two methods can reconstruct the high-resolution flow field in just a few seconds,and further greatlyenrich the application of high-resolution reconstruction technology in fluid instability.Compared with the ordinaryconvolutional neural network,the multi-time-path convolutional neural network model has smaller error and canrestore more details of the flow field.The influence of low-resolution flow field data obtained by the two poolingmethods on the convolutional neural networks model is also discussed.展开更多
Microseismic monitoring system is one of the effective methods for deep mining geo-stress monitoring.The principle of microseismic monitoring system is to analyze the mechanical parameters contained in microseismic ev...Microseismic monitoring system is one of the effective methods for deep mining geo-stress monitoring.The principle of microseismic monitoring system is to analyze the mechanical parameters contained in microseismic events for providing accurate information of rockmass.The accurate identification of microseismic events and blasts determines the timeliness and accuracy of early warning of microseismic monitoring technology.An image identification model based on Convolutional Neural Network(CNN)is established in this paper for the seismic waveforms of microseismic events and blasts.Firstly,the training set,test set,and validation set are collected,which are composed of 5250,1500,and 750 seismic waveforms of microseismic events and blasts,respectively.The classified data sets are preprocessed and input into the constructed CNN in CPU mode for training.Results show that the accuracies of microseismic events and blasts are 99.46%and 99.33%in the test set,respectively.The accuracies of microseismic events and blasts are 100%and 98.13%in the validation set,respectively.The proposed method gives superior performance when compared with existed methods.The accuracies of models using logistic regression and artificial neural network(ANN)based on the same data set are 54.43%and 67.9%in the test set,respectively.Then,the ROC curves of the three models are obtained and compared,which show that the CNN gives an absolute advantage in this classification model when the original seismic waveform are used in training the model.It not only decreases the influence of individual differences in experience,but also removes the errors induced by source and waveform parameters.It is proved that the established discriminant method improves the efficiency and accuracy of microseismic data processing for monitoring rock instability and seismicity.展开更多
Energy expenditure estimation can be used to measure the exercise load and physical condition of different individuals, such as soldiers, athletes, firemen, etc., during their training and work. Energy expenditure est...Energy expenditure estimation can be used to measure the exercise load and physical condition of different individuals, such as soldiers, athletes, firemen, etc., during their training and work. Energy expenditure estimation methods based on computer vision have rapidly developed in recent years. Compared with sensor-based methods, such methods are capable of monitoring several target persons at the same time, and the subjects do not need to wear different sensor devices that hamper their movement. In this paper, we propose a cross-attention spatial–temporal convolutional neural network to predict the energy expenditure of people under different exercise intensities. The model explores the relationship between changes in the human skeleton and energy expenditure intensity. In addition, a cross-attention correction module is used to reduce the negative effects of individual physical fitness characteristics during energy expenditure estimation. The experimental results show that our proposed method achieves high accuracy for energy expenditure estimation and performs better than existing computer vision-based energy expenditure estimation methods do. The proposed method can be widely used in various physical activity scenarios to measure energy expenditure, increasing the convenience of usage.展开更多
Considering the problem that the scattering echo images of airborne Doppler weather radar are often reduced by ground clutters,the accuracy and confidence of meteorology target detection are reduced.In this paper,a de...Considering the problem that the scattering echo images of airborne Doppler weather radar are often reduced by ground clutters,the accuracy and confidence of meteorology target detection are reduced.In this paper,a deep convolutional neural network(DCNN)is proposed for meteorology target detection and ground clutter suppression with a large collection of airborne weather radar images as network input.For each weather radar image,the corresponding digital elevation model(DEM)image is extracted on basis of the radar antenna scan-ning parameters and plane position,and is further fed to the net-work as a supplement for ground clutter suppression.The fea-tures of actual meteorology targets are learned in each bottle-neck module of the proposed network and convolved into deeper iterations in the forward propagation process.Then the network parameters are updated by the back propagation itera-tion of the training error.Experimental results on the real mea-sured images show that our proposed DCNN outperforms the counterparts in terms of six evaluation factors.Meanwhile,the network outputs are in good agreement with the expected mete-orology detection results(labels).It is demonstrated that the pro-posed network would have a promising meteorology observa-tion application with minimal effort on network variables or parameter changes.展开更多
This paper concerns the problem of object segmentation in real-time for picking system. A region proposal method inspired by human glance based on the convolutional neural network is proposed to select promising regio...This paper concerns the problem of object segmentation in real-time for picking system. A region proposal method inspired by human glance based on the convolutional neural network is proposed to select promising regions, allowing more processing is reserved only for these regions. The speed of object segmentation is significantly improved by the region proposal method.By the combination of the region proposal method based on the convolutional neural network and superpixel method, the category and location information can be used to segment objects and image redundancy is significantly reduced. The processing time is reduced considerably by this to achieve the real time. Experiments show that the proposed method can segment the interested target object in real time on an ordinary laptop.展开更多
Weakly supervised semantic segmentation(WSSS)is a tricky task,which only provides category information for segmentation prediction.Thus,the key stage of WSSS is to generate the pseudo labels.For convolutional neural n...Weakly supervised semantic segmentation(WSSS)is a tricky task,which only provides category information for segmentation prediction.Thus,the key stage of WSSS is to generate the pseudo labels.For convolutional neural network(CNN)based methods,in which class activation mapping(CAM)is proposed to obtain the pseudo labels,and only concentrates on the most discriminative parts.Recently,transformer-based methods utilize attention map from the multi-headed self-attention(MHSA)module to predict pseudo labels,which usually contain obvious background noise and incoherent object area.To solve the above problems,we use the Conformer as our backbone,which is a parallel network based on convolutional neural network(CNN)and Transformer.The two branches generate pseudo labels and refine them independently,and can effectively combine the advantages of CNN and Transformer.However,the parallel structure is not close enough in the information communication.Thus,parallel structure can result in poor details about pseudo labels,and the background noise still exists.To alleviate this problem,we propose enhancing convolution CAM(ECCAM)model,which have three improved modules based on enhancing convolution,including deeper stem(DStem),convolutional feed-forward network(CFFN)and feature coupling unit with convolution(FCUConv).The ECCAM could make Conformer have tighter interaction between CNN and Transformer branches.After experimental verification,the improved modules we propose can help the network perceive more local information from images,making the final segmentation results more refined.Compared with similar architecture,our modules greatly improve the semantic segmentation performance and achieve70.2%mean intersection over union(mIoU)on the PASCAL VOC 2012 dataset.展开更多
Taking Jiuhong Modern Agriculture Demonstration Park of Heilongjiang Province as the base for rice disease image acquisition,a total of 841 images of the four different diseases,including rice blast,stripe leaf blight...Taking Jiuhong Modern Agriculture Demonstration Park of Heilongjiang Province as the base for rice disease image acquisition,a total of 841 images of the four different diseases,including rice blast,stripe leaf blight,red blight and bacterial brown spot,were obtained.In this study,an interleaved attention neural network(IANN)was proposed to realize the recognition of rice disease images and an interleaved group convolutions(IGC)network was introduced to reduce the number of convolutional parameters,which realized the information interaction between channels.Based on the convolutional block attention module(CBAM),attention was paid to the features of results of the primary group convolution in the cross-group convolution to improve the classification performance of the deep learning model.The results showed that the classification accuracy of IANN was 96.14%,which was 4.72%higher than that of the classical convolutional neural network(CNN).This study showed a new idea for the efficient training of neural networks in the case of small samples and provided a reference for the image recognition and diagnosis of rice and other crop diseases.展开更多
Semantic segmentation of eye images is a complex task with important applications in human–computer interaction,cognitive science,and neuroscience.Achieving real-time,accurate,and robust segmentation algorithms is cr...Semantic segmentation of eye images is a complex task with important applications in human–computer interaction,cognitive science,and neuroscience.Achieving real-time,accurate,and robust segmentation algorithms is crucial for computationally limited portable devices such as augmented reality and virtual reality.With the rapid advancements in deep learning,many network models have been developed specifically for eye image segmentation.Some methods divide the segmentation process into multiple stages to achieve model parameter miniaturization while enhancing output through post processing techniques to improve segmentation accuracy.These approaches significantly increase the inference time.Other networks adopt more complex encoding and decoding modules to achieve end-to-end output,which requires substantial computation.Therefore,balancing the model’s size,accuracy,and computational complexity is essential.To address these challenges,we propose a lightweight asymmetric UNet architecture and a projection loss function.We utilize ResNet-3 layer blocks to enhance feature extraction efficiency in the encoding stage.In the decoding stage,we employ regular convolutions and skip connections to upscale the feature maps from the latent space to the original image size,balancing the model size and segmentation accuracy.In addition,we leverage the geometric features of the eye region and design a projection loss function to further improve the segmentation accuracy without adding any additional inference computational cost.We validate our approach on the OpenEDS2019 dataset for virtual reality and achieve state-of-the-art performance with 95.33%mean intersection over union(mIoU).Our model has only 0.63M parameters and 350 FPS,which are 68%and 200%of the state-of-the-art model RITNet,respectively.展开更多
Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devi...Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devices have made power load data increasingly complex and volatile.This places higher demands on the prediction and analysis of power loads.In order to improve the prediction accuracy of short-term power load,a CNN-BiLSTMTPA short-term power prediction model based on the Improved Whale Optimization Algorithm(IWOA)with mixed strategies was proposed.Firstly,the model combined the Convolutional Neural Network(CNN)with the Bidirectional Long Short-Term Memory Network(BiLSTM)to fully extract the spatio-temporal characteristics of the load data itself.Then,the Temporal Pattern Attention(TPA)mechanism was introduced into the CNN-BiLSTM model to automatically assign corresponding weights to the hidden states of the BiLSTM.This allowed the model to differentiate the importance of load sequences at different time intervals.At the same time,in order to solve the problem of the difficulties of selecting the parameters of the temporal model,and the poor global search ability of the whale algorithm,which is easy to fall into the local optimization,the whale algorithm(IWOA)was optimized by using the hybrid strategy of Tent chaos mapping and Levy flight strategy,so as to better search the parameters of the model.In this experiment,the real load data of a region in Zhejiang was taken as an example to analyze,and the prediction accuracy(R2)of the proposed method reached 98.83%.Compared with the prediction models such as BP,WOA-CNN-BiLSTM,SSA-CNN-BiLSTM,CNN-BiGRU-Attention,etc.,the experimental results showed that the model proposed in this study has a higher prediction accuracy.展开更多
This paper presents the application of a novel AI-based approach,Neural Physics,to produce high-fidelity simulations of train aerodynamics.Neural Physics is built upon convolutional neural networks(CNNs),where the wei...This paper presents the application of a novel AI-based approach,Neural Physics,to produce high-fidelity simulations of train aerodynamics.Neural Physics is built upon convolutional neural networks(CNNs),where the weights are explicitly determined by classical numerical discretisation schemes rather than by training.By leveraging the power of AI technology,this recent approach results in code that can run easily on GPUs and AI processors,achieving high computational speed without sacrificing accuracy.The approach uses an implicit large eddy simulation method based on a non-linear Petrov-Galerkin method to model the unresolved turbulence.Furthermore,for higher-order finite elements,the convolutional finite element method(ConvFEM)is used,which greatly simplifies the implementation of higher-order elements within the NN 4 DPEs approach.We demonstrate the capability of Neural Physics by simulating a freight Locomotive Class 66 and a partially loaded freight train operating in an open field environment with and without cross wind.This is the first time that ConvFEM has been applied to high-speed fluid flow problems in complex geometries.The results are validated against existing numerical results and experimental measurements,and show good agreement in terms of pressure and velocity distributions around the train body.展开更多
The fifth-generation (5G) communication requires a highly accurate estimation of the channel state information (CSI)to take advantage of the massive multiple-input multiple-output(MIMO) system. However, traditional ch...The fifth-generation (5G) communication requires a highly accurate estimation of the channel state information (CSI)to take advantage of the massive multiple-input multiple-output(MIMO) system. However, traditional channel estimation methods do not always yield reliable estimates. The methodology of this paper consists of deep residual shrinkage network (DRSN)neural network-based method that is used to solve this problem.Thus, the channel estimation approach, based on DRSN with its learning ability of noise-containing data, is first introduced. Then,the DRSN is used to train the noise reduction process based on the results of the least square (LS) channel estimation while applying the pilot frequency subcarriers, where the initially estimated subcarrier channel matrix is considered as a three-dimensional tensor of the DRSN input. Afterward, a mixed signal to noise ratio (SNR) training data strategy is proposed based on the learning ability of DRSN under different SNRs. Moreover, a joint mixed scenario training strategy is carried out to test the multi scenarios robustness of DRSN. As for the findings, the numerical results indicate that the DRSN method outperforms the spatial-frequency-temporal convolutional neural networks (SF-CNN)with similar computational complexity and achieves better advantages in the full SNR range than the minimum mean squared error (MMSE) estimator with a limited dataset. Moreover, the DRSN approach shows robustness in different propagation environments.展开更多
针对景区手写诗词存在背景纹理复杂、字体尺寸及风格多样等特点导致景区游客难以识别手写诗词的问题,首先,分析研究景区手写诗词的识别场景,设计景区诗词检测网络(detection of poetry in scenic areas-network,DPSA-Net)以提取景区手...针对景区手写诗词存在背景纹理复杂、字体尺寸及风格多样等特点导致景区游客难以识别手写诗词的问题,首先,分析研究景区手写诗词的识别场景,设计景区诗词检测网络(detection of poetry in scenic areas-network,DPSA-Net)以提取景区手写诗词不同尺度的特征,并结合手写诗词字符间的链接依赖关系实现景区手写诗词检测;其次,设计了卷积循环聚合网络(convolution recurrent aggregation network,CRA-Net)以对景区手写诗词进行识别,结合卷积神经网络(convolutional neural networks,CNN)和双向长短期记忆网络提取手写诗词图像的序列特征,并通过聚合交叉熵(aggregation cross-entropy,ACE)实现特征向文本的转换;最后,结合景区知识图谱对CRA-Net的输出进行校正,进而提高景区手写诗词的识别准确率。实验结果表明,通过景区手写诗词矫正技术对CRA-Net的识别结果矫正后,识别准确率达到了79.04%,同时,该技术具有较好的抗干扰能力和良好的应用前景。展开更多
文摘Human disturbance activities is one of the main reasons for inducing geohazards.Ecological impact assessment metrics of roads are inconsistent criteria and multiple.From the perspective of visual observation,the environment damage can be shown through detecting the uncovered area of vegetation in the images along road.To realize this,an end-to-end environment damage detection model based on convolutional neural network is proposed.A 50-layer residual network is used to extract feature map.The initial parameters are optimized by transfer learning.An example is shown by this method.The dataset including cliff and landslide damage are collected by us along road in Shennongjia national forest park.Results show 0.4703 average precision(AP)rating for cliff damage and 0.4809 average precision(AP)rating for landslide damage.Compared with YOLOv3,our model shows a better accuracy in cliff and landslide detection although a certain amount of speed is sacrificed.
基金supported by the National Natural Science Foundation of China (61471021)。
文摘Non-orthogonal multiple access(NOMA), featuring high spectrum efficiency, massive connectivity and low latency, holds immense potential to be a novel multi-access technique in fifth-generation(5G) communication. Successive interference cancellation(SIC) is proved to be an effective method to detect the NOMA signal by ordering the power of received signals and then decoding them. However, the error accumulation effect referred to as error propagation is an inevitable problem. In this paper,we propose a convolutional neural networks(CNNs) approach to restore the desired signal impaired by the multiple input multiple output(MIMO) channel. Especially in the uplink NOMA scenario,the proposed method can decode multiple users' information in a cluster instantaneously without any traditional communication signal processing steps. Simulation experiments are conducted in the Rayleigh channel and the results demonstrate that the error performance of the proposed learning system outperforms that of the classic SIC detection. Consequently, deep learning has disruptive potential to replace the conventional signal detection method.
基金National Natural Science Foundation of China(1180500311947102+4 种基金12004005)Natural Science Foundation of Anhui Province(2008085MA162008085QA26)University Synergy Innovation Program of Anhui Province(GXXT-2022-039)State Key Laboratory of Advanced Electromagnetic Technology(Grant No.AET 2024KF006)。
文摘High-resolution flow field data has important applications in meteorology,aerospace engineering,high-energy physics and other fields.Experiments and numerical simulations are two main ways to obtain high-resolution flow field data,while the high experiment cost and computing resources for simulation hinder the specificanalysis of flow field evolution.With the development of deep learning technology,convolutional neural networks areused to achieve high-resolution reconstruction of the flow field.In this paper,an ordinary convolutional neuralnetwork and a multi-time-path convolutional neural network are established for the ablative Rayleigh-Taylorinstability.These two methods can reconstruct the high-resolution flow field in just a few seconds,and further greatlyenrich the application of high-resolution reconstruction technology in fluid instability.Compared with the ordinaryconvolutional neural network,the multi-time-path convolutional neural network model has smaller error and canrestore more details of the flow field.The influence of low-resolution flow field data obtained by the two poolingmethods on the convolutional neural networks model is also discussed.
基金Projects(51822407,51774327,51664016)supported by the National Natural Science Foundation of China。
文摘Microseismic monitoring system is one of the effective methods for deep mining geo-stress monitoring.The principle of microseismic monitoring system is to analyze the mechanical parameters contained in microseismic events for providing accurate information of rockmass.The accurate identification of microseismic events and blasts determines the timeliness and accuracy of early warning of microseismic monitoring technology.An image identification model based on Convolutional Neural Network(CNN)is established in this paper for the seismic waveforms of microseismic events and blasts.Firstly,the training set,test set,and validation set are collected,which are composed of 5250,1500,and 750 seismic waveforms of microseismic events and blasts,respectively.The classified data sets are preprocessed and input into the constructed CNN in CPU mode for training.Results show that the accuracies of microseismic events and blasts are 99.46%and 99.33%in the test set,respectively.The accuracies of microseismic events and blasts are 100%and 98.13%in the validation set,respectively.The proposed method gives superior performance when compared with existed methods.The accuracies of models using logistic regression and artificial neural network(ANN)based on the same data set are 54.43%and 67.9%in the test set,respectively.Then,the ROC curves of the three models are obtained and compared,which show that the CNN gives an absolute advantage in this classification model when the original seismic waveform are used in training the model.It not only decreases the influence of individual differences in experience,but also removes the errors induced by source and waveform parameters.It is proved that the established discriminant method improves the efficiency and accuracy of microseismic data processing for monitoring rock instability and seismicity.
基金supported by the National Key Research and Development Program of China (Grant No.2022YFC3600403)Emerging Interdisciplinary Platform for Medicine and Engineering in Sports (EIPMES),Beijing,China。
文摘Energy expenditure estimation can be used to measure the exercise load and physical condition of different individuals, such as soldiers, athletes, firemen, etc., during their training and work. Energy expenditure estimation methods based on computer vision have rapidly developed in recent years. Compared with sensor-based methods, such methods are capable of monitoring several target persons at the same time, and the subjects do not need to wear different sensor devices that hamper their movement. In this paper, we propose a cross-attention spatial–temporal convolutional neural network to predict the energy expenditure of people under different exercise intensities. The model explores the relationship between changes in the human skeleton and energy expenditure intensity. In addition, a cross-attention correction module is used to reduce the negative effects of individual physical fitness characteristics during energy expenditure estimation. The experimental results show that our proposed method achieves high accuracy for energy expenditure estimation and performs better than existing computer vision-based energy expenditure estimation methods do. The proposed method can be widely used in various physical activity scenarios to measure energy expenditure, increasing the convenience of usage.
基金supported by the China Ministry of Industry and Information Technology Foundation and Aeronautical Science Foundation of China(ASFC-201920007002)the National Key Research and Development Plan(2021YFB1600603)the Open Fund of Key Laboratory of Civil Aircraft Airworthiness Technology,Civil Aviation University of China.
文摘Considering the problem that the scattering echo images of airborne Doppler weather radar are often reduced by ground clutters,the accuracy and confidence of meteorology target detection are reduced.In this paper,a deep convolutional neural network(DCNN)is proposed for meteorology target detection and ground clutter suppression with a large collection of airborne weather radar images as network input.For each weather radar image,the corresponding digital elevation model(DEM)image is extracted on basis of the radar antenna scan-ning parameters and plane position,and is further fed to the net-work as a supplement for ground clutter suppression.The fea-tures of actual meteorology targets are learned in each bottle-neck module of the proposed network and convolved into deeper iterations in the forward propagation process.Then the network parameters are updated by the back propagation itera-tion of the training error.Experimental results on the real mea-sured images show that our proposed DCNN outperforms the counterparts in terms of six evaluation factors.Meanwhile,the network outputs are in good agreement with the expected mete-orology detection results(labels).It is demonstrated that the pro-posed network would have a promising meteorology observa-tion application with minimal effort on network variables or parameter changes.
基金supported by the National Natural Science Foundation of China(61233010 61305106)+2 种基金the Shanghai Natural Science Foundation(17ZR1409700 18ZR1415300)the basic research project of Shanghai Municipal Science and Technology Commission(16JC1400900)
文摘This paper concerns the problem of object segmentation in real-time for picking system. A region proposal method inspired by human glance based on the convolutional neural network is proposed to select promising regions, allowing more processing is reserved only for these regions. The speed of object segmentation is significantly improved by the region proposal method.By the combination of the region proposal method based on the convolutional neural network and superpixel method, the category and location information can be used to segment objects and image redundancy is significantly reduced. The processing time is reduced considerably by this to achieve the real time. Experiments show that the proposed method can segment the interested target object in real time on an ordinary laptop.
文摘Weakly supervised semantic segmentation(WSSS)is a tricky task,which only provides category information for segmentation prediction.Thus,the key stage of WSSS is to generate the pseudo labels.For convolutional neural network(CNN)based methods,in which class activation mapping(CAM)is proposed to obtain the pseudo labels,and only concentrates on the most discriminative parts.Recently,transformer-based methods utilize attention map from the multi-headed self-attention(MHSA)module to predict pseudo labels,which usually contain obvious background noise and incoherent object area.To solve the above problems,we use the Conformer as our backbone,which is a parallel network based on convolutional neural network(CNN)and Transformer.The two branches generate pseudo labels and refine them independently,and can effectively combine the advantages of CNN and Transformer.However,the parallel structure is not close enough in the information communication.Thus,parallel structure can result in poor details about pseudo labels,and the background noise still exists.To alleviate this problem,we propose enhancing convolution CAM(ECCAM)model,which have three improved modules based on enhancing convolution,including deeper stem(DStem),convolutional feed-forward network(CFFN)and feature coupling unit with convolution(FCUConv).The ECCAM could make Conformer have tighter interaction between CNN and Transformer branches.After experimental verification,the improved modules we propose can help the network perceive more local information from images,making the final segmentation results more refined.Compared with similar architecture,our modules greatly improve the semantic segmentation performance and achieve70.2%mean intersection over union(mIoU)on the PASCAL VOC 2012 dataset.
基金Supported by the Heilongjiang Provincial Key Research and Development Program Guidance Project(GZ20210103)。
文摘Taking Jiuhong Modern Agriculture Demonstration Park of Heilongjiang Province as the base for rice disease image acquisition,a total of 841 images of the four different diseases,including rice blast,stripe leaf blight,red blight and bacterial brown spot,were obtained.In this study,an interleaved attention neural network(IANN)was proposed to realize the recognition of rice disease images and an interleaved group convolutions(IGC)network was introduced to reduce the number of convolutional parameters,which realized the information interaction between channels.Based on the convolutional block attention module(CBAM),attention was paid to the features of results of the primary group convolution in the cross-group convolution to improve the classification performance of the deep learning model.The results showed that the classification accuracy of IANN was 96.14%,which was 4.72%higher than that of the classical convolutional neural network(CNN).This study showed a new idea for the efficient training of neural networks in the case of small samples and provided a reference for the image recognition and diagnosis of rice and other crop diseases.
文摘现有的基于卷积神经网络(convolutional neural network,CNN)的环路滤波器倾向于将多个网络应用于不同的量化参数(quantization parameter,QP),消耗训练模型中的大量资源,并增加内存负担。针对这一问题,提出一种基于CNN的QP自适应环路滤波器。首先,设计一个轻量级分类网络,按照滤波难易程度将编码树单元(coding tree unit,CTU)划分为难、中、易3类;然后,构建3个融合了特征信息增强融合模块的基于CNN的滤波网络,以满足不同QP下的3类CTU滤波需求。将所提出的环路滤波器集成到多功能视频编码(versatile video coding,VVC)标准H.266/VVC的测试软件VTM 6.0中,替换原有的去块效应滤波器(deblocking filter,DBF)、样本自适应偏移(sample adaptive offset,SAO)滤波器和自适应环路滤波器。实验结果表明,该方法平均降低了3.14%的比特率差值(Bjøntegaard delta bit rate,BD-BR),与其他基于CNN的环路滤波器相比,显著提高了压缩效率,并减少了压缩伪影。
基金supported by the HFIPS Director’s Foundation(YZJJ202207-TS),the National Natural Science Foundation of China(82371931)the Natural Science Foundation of Anhui Province(2008085MC69)+3 种基金the Natural Science Foundation of Hefei City(2021033)the General Scientific Research Project of Anhui Provincial Health Commission(AHWJ2021b150)the Collaborative Innovation Program of Hefei Science Center,CAS(2021HSC-CIP013)the Anhui Province Key Research and Development Project(202204295107020004).
文摘Semantic segmentation of eye images is a complex task with important applications in human–computer interaction,cognitive science,and neuroscience.Achieving real-time,accurate,and robust segmentation algorithms is crucial for computationally limited portable devices such as augmented reality and virtual reality.With the rapid advancements in deep learning,many network models have been developed specifically for eye image segmentation.Some methods divide the segmentation process into multiple stages to achieve model parameter miniaturization while enhancing output through post processing techniques to improve segmentation accuracy.These approaches significantly increase the inference time.Other networks adopt more complex encoding and decoding modules to achieve end-to-end output,which requires substantial computation.Therefore,balancing the model’s size,accuracy,and computational complexity is essential.To address these challenges,we propose a lightweight asymmetric UNet architecture and a projection loss function.We utilize ResNet-3 layer blocks to enhance feature extraction efficiency in the encoding stage.In the decoding stage,we employ regular convolutions and skip connections to upscale the feature maps from the latent space to the original image size,balancing the model size and segmentation accuracy.In addition,we leverage the geometric features of the eye region and design a projection loss function to further improve the segmentation accuracy without adding any additional inference computational cost.We validate our approach on the OpenEDS2019 dataset for virtual reality and achieve state-of-the-art performance with 95.33%mean intersection over union(mIoU).Our model has only 0.63M parameters and 350 FPS,which are 68%and 200%of the state-of-the-art model RITNet,respectively.
文摘Load forecasting is of great significance to the development of new power systems.With the advancement of smart grids,the integration and distribution of distributed renewable energy sources and power electronics devices have made power load data increasingly complex and volatile.This places higher demands on the prediction and analysis of power loads.In order to improve the prediction accuracy of short-term power load,a CNN-BiLSTMTPA short-term power prediction model based on the Improved Whale Optimization Algorithm(IWOA)with mixed strategies was proposed.Firstly,the model combined the Convolutional Neural Network(CNN)with the Bidirectional Long Short-Term Memory Network(BiLSTM)to fully extract the spatio-temporal characteristics of the load data itself.Then,the Temporal Pattern Attention(TPA)mechanism was introduced into the CNN-BiLSTM model to automatically assign corresponding weights to the hidden states of the BiLSTM.This allowed the model to differentiate the importance of load sequences at different time intervals.At the same time,in order to solve the problem of the difficulties of selecting the parameters of the temporal model,and the poor global search ability of the whale algorithm,which is easy to fall into the local optimization,the whale algorithm(IWOA)was optimized by using the hybrid strategy of Tent chaos mapping and Levy flight strategy,so as to better search the parameters of the model.In this experiment,the real load data of a region in Zhejiang was taken as an example to analyze,and the prediction accuracy(R2)of the proposed method reached 98.83%.Compared with the prediction models such as BP,WOA-CNN-BiLSTM,SSA-CNN-BiLSTM,CNN-BiGRU-Attention,etc.,the experimental results showed that the model proposed in this study has a higher prediction accuracy.
基金Projects(EPSRC EP/Y005732/1,EP/Y018680/1,EP/T003189/1,EP/V040235/1,EP/Y024257/1 and EP/T000414/1)supported by UK Research and Innovation(UKRI),UKProject(APP44894/UKRI 1281)supported by UKRI councils(NERC,AHRC,ESRC,MRC and DEFRA),UK。
文摘This paper presents the application of a novel AI-based approach,Neural Physics,to produce high-fidelity simulations of train aerodynamics.Neural Physics is built upon convolutional neural networks(CNNs),where the weights are explicitly determined by classical numerical discretisation schemes rather than by training.By leveraging the power of AI technology,this recent approach results in code that can run easily on GPUs and AI processors,achieving high computational speed without sacrificing accuracy.The approach uses an implicit large eddy simulation method based on a non-linear Petrov-Galerkin method to model the unresolved turbulence.Furthermore,for higher-order finite elements,the convolutional finite element method(ConvFEM)is used,which greatly simplifies the implementation of higher-order elements within the NN 4 DPEs approach.We demonstrate the capability of Neural Physics by simulating a freight Locomotive Class 66 and a partially loaded freight train operating in an open field environment with and without cross wind.This is the first time that ConvFEM has been applied to high-speed fluid flow problems in complex geometries.The results are validated against existing numerical results and experimental measurements,and show good agreement in terms of pressure and velocity distributions around the train body.
基金supported by the National Key Scientific Instrument and Equipment Development Project(61827801).
文摘The fifth-generation (5G) communication requires a highly accurate estimation of the channel state information (CSI)to take advantage of the massive multiple-input multiple-output(MIMO) system. However, traditional channel estimation methods do not always yield reliable estimates. The methodology of this paper consists of deep residual shrinkage network (DRSN)neural network-based method that is used to solve this problem.Thus, the channel estimation approach, based on DRSN with its learning ability of noise-containing data, is first introduced. Then,the DRSN is used to train the noise reduction process based on the results of the least square (LS) channel estimation while applying the pilot frequency subcarriers, where the initially estimated subcarrier channel matrix is considered as a three-dimensional tensor of the DRSN input. Afterward, a mixed signal to noise ratio (SNR) training data strategy is proposed based on the learning ability of DRSN under different SNRs. Moreover, a joint mixed scenario training strategy is carried out to test the multi scenarios robustness of DRSN. As for the findings, the numerical results indicate that the DRSN method outperforms the spatial-frequency-temporal convolutional neural networks (SF-CNN)with similar computational complexity and achieves better advantages in the full SNR range than the minimum mean squared error (MMSE) estimator with a limited dataset. Moreover, the DRSN approach shows robustness in different propagation environments.
