检索结果-南通市图书馆

Data augmentation via joint multi-scale CNN and multi-channel attention for bumblebee image generation

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

The Journal of China Universities of Posts and Telecommunications 2023年第3期30卷 32-40,98页

作者： Du Rong Chen Shudong Li Weiwei Zhang Xueting Wang Xianhui Ge Jin Intelligent Manufacturing Electronics Research and Development Center Institute of Microelectronics of the Chinese Academy of SciencesBeijing 100029China School of Integrated Circuits University of Chinese Academy of SciencesBeijing 100049China Institute of Zoology Chinese Academy of SciencesBeijing 100080China Center for Excellence in Biotic Interactions University of Chinese Academy of SciencesBeijing 100049China

The difficulty of bumblebee data collecting and the laborious nature of bumblebee data annotation sometimes result in a lack of training data,which impairs the effectiveness of deep learning based counting methods.Given that it is challenging to produce the detailed background information in the generated bumblebee images using current data augmentation methods,in this paper,a joint multi-scale convolutional neural network and multi-channel attention based generative adversarial network(MMGAN)is proposed.MMGAN generates the bumblebee image in accordance with the corresponding density map marking the bumblebee positions.Specifically,the multi-scale convolutional neural network(CNN)module utilizes multiple convolution kernels to completely extract features of different scales from the input bumblebee image and density map.To generate various targets in the generated image,the multi-channel attention module builds numerous intermediate generation layers and attention maps.These targets are then stacked to produce a bumblebee image with a specific number of bumblebees.The proposed model obtains the greatest performance in bumblebee image generating tasks,and such generated bumblebee images considerably improve the efficiency of deep learning based counting methods in bumblebee counting applications.

关键词： data augmentation image generation attention mechanism

A Spatial-Temporal attention Model for Human Trajectory Prediction

在线全文

学校读者我要写书评

暂无评论

IEEE/CAA Journal of Automatica Sinica 2020年第4期7卷 965-974页

作者： Xiaodong Zhao Yaran Chen Jin Guo Dongbin Zhao the School of Automation and Electrical Engineering University of Science and Technology BeijingBeijing 100083 the State Key Laboratory of Management and Control for Complex Systems Institute of AutomationChinese Academy of SciencesBeijing 100190China the Key Laboratory of Knowledge Automation for Industrial Processes Ministry of EducationBeijing 100083China

Human trajectory prediction is essential and promising in many related applications. This is challenging due to the uncertainty of human behaviors, which can be influenced not only by himself, but also by the surrounding environment. Recent works based on long-short term memory(LSTM) models have brought tremendous improvements on the task of trajectory prediction. However, most of them focus on the spatial influence of humans but ignore the temporal influence. In this paper, we propose a novel spatial-temporal attention(ST-attention) model,which studies spatial and temporal affinities jointly. Specifically,we introduce an attention mechanism to extract temporal affinity,learning the importance for historical trajectory information at different time instants. To explore spatial affinity, a deep neural network is employed to measure different importance of the neighbors. Experimental results show that our method achieves competitive performance compared with state-of-the-art methods on publicly available datasets.

关键词： attention mechanism long-short term memory(LSTM) spatial-temporal model trajectory prediction

Algorithm of Helmet Wearing Detection Based on AT-YOLO Deep Mode

在线全文

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2021年第10期69卷 159-174页

作者： Qingyang Zhou Jiaohua Qin Xuyu Xiang Yun Tan Neal NXiong College of Computer Science and Information Technology Central South University of Forestry&TechnologyChangsha410004China Department of Mathematics and Computer Science Northeastern State UniversityTahlequah74464OKUSA

The existing safety helmet detection methods are mainly based on one-stage object detection algorithms with high detection speed to reach the real-time detection requirements,but they can’t accurately detect small objects and objects with obstructions.Therefore,we propose a helmet detection algorithm based on the attention mechanism(AT-YOLO).First of all,a channel attention module is added to the YOLOv3 backbone network,which can adaptively calibrate the channel features of the direction to improve the feature utilization,and a spatial attention module is added to the neck of the YOLOv3 network to capture the correlation between any positions in the feature map so that to increase the receptive field of the network.Secondly,we use DIoU(Distance Intersection over Union)bounding box regression loss function,it not only improving the measurement of bounding box regression loss but also increases the normalized distance loss between the prediction boxes and the target boxes,which makes the network more accurate in detecting small objects and faster in convergence.Finally,we explore the training strategy of the network model,which improves network performance without increasing the inference cost.Experiments show that the mAP of the proposed method reaches 96.5%,and the detection speed can reach 27 fps.Compared with other existing methods,it has better performance in detection accuracy and speed.

关键词： Safety helmet detection attention mechanism convolutional neural network training strategies

Straw Segmentation Algorithm Based on Modified UNet in Complex Farmland Environment

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2021年第1期66卷 247-262页

作者： Yuanyuan Liu Shuo Zhang Haiye Yu Yueyong Wang Yuehan Feng Jiahui Sun Xiaokang Zhou College of Information Technology Jilin Agriculture UniversityChangchun130118China Department of Biosystems and Agricultural Engineering Oklahoma State UniversityStillwater74078USA Key Laboratory of Bionic Engineering Ministry of EducationJilin UniversityChangchun130025China College of Engineering and Technology Jilin Agricultural UniversityChangchun130118China

Intelligent straw coverage detection plays an important role in agricultural production and the ecological environment.Traditional pattern recognition has some problems,such as low precision and a long processing time,when segmenting complex farmland,which cannot meet the conditions of embedded equipment deployment.Based on these problems,we proposed a novel deep learning model with high accuracy,small model size and fast running speed named Residual Unet with attention mechanism using depthwise convolution(RADw–UNet).This algorithm is based on the UNet symmetric codec model.All the feature extraction modules of the network adopt the residual structure,and the whole network only adopts 8 times the downsampling rate to reduce the redundant parameters.To better extract the semantic information of the spatial and channel dimensions,the depthwise convolutional residual block is designed to be used in feature maps with larger depths to reduce the number of parameters while improving the model accuracy.Meanwhile,the multi–level attention mechanism is introduced in the skip connection to effectively integrate the information of the low–level and high–level feature maps.The experimental results showed that the segmentation performance of RADw–UNet outperformed traditional methods and the UNet algorithm.The algorithm achieved an mIoU of 94.9%,the number of trainable parameters was only approximately 0.26 M,and the running time for a single picture was less than 0.03 s.

关键词： Straw segmentation convolutional neural network residual structure depthwise convolution attention mechanism

An attention-Based Friend Recommendation Model in Social Network

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2020年第12期65卷 2475-2488页

作者： Chongchao Cai Huahu Xu Jie Wan Baiqing Zhou Xiongwei Xie School of Computer Engineering and Science Shanghai UniversityShanghai200444China College of Logistic and Information Engineering Huzhou Vocational&Technical CollegeHuzhou313000China Google New YorkUSA

In social networks,user attention affects the user’s decision-making,resulting in a performance alteration of the recommendation systems.Existing systems make recommendations mainly according to users’preferences with a particular focus on items.However,the significance of users’attention and the difference in the influence of different users and items are often ignored.Thus,this paper proposes an attention-based multi-layer friend recommendation model to mitigate information overload in social networks.We first constructed the basic user and item matrix via convolutional neural networks(CNN).Then,we obtained user preferences by using the relationships between users and items,which were later inputted into our model to learn the preferences between friends.The error performance of the proposed method was compared with the traditional solutions based on collaborative filtering.A comprehensive performance evaluation was also conducted using large-scale real-world datasets collected from three popular location-based social networks.The experimental results revealed that our proposal outperforms the traditional methods in terms of recommendation performance.

关键词： Friend recommendation collaborative filtering attention mechanism deep learning

End-to-end encrypted network traffic classification method based on deep learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

The Journal of China Universities of Posts and Telecommunications 2020年第3期27卷 21-30页

作者： Tian Shiming Gong Feixiang Mo Shuang Li Meng Wu Wenrui Xiao Ding Department of Power Consumption and Energy Efficiency China Electric Power Research Institute Co.Ltd.Beijing 100192China School of Com puter Science Beijing University of Posts and Telecom municationsBeijing 100876China State Grid Corporation of China Beijing 100031China

Network traffic classification,which matches network traffic for a specific class of different granularities,plays a vital role in the domain of network administration and cyber security.With the rapid development of network communication techniques,more and more network applications adopt encryption techniques during communication,which brings significant challenges to traditional network traffic classification methods.On the one hand,traditional methods mainly depend on matching features on the application layer of the ISO/OSI reference model,which leads to the failure of classifying encrypted traffic.On the other hand,machine learning-based methods require human-made features from network traffic data by human experts,which renders it difficult for them to deal with complex network protocols.In this paper,the convolution attention network(CAT)is proposed to overcom those difficulties.As an end-to-end model,CAT takes raw data as input and returns classification results automatically,with engineering by human experts.In CAT,firstly,the importance of different bytes with an attention mechanism of network traffic is achieved.Then,convolution neural network(CNN)is used to learn features automatically and feed the output into a softmax function to get classification results.It enables CAT to learn enough information from network traffic data and ensure the classified accuracy.Extensive experiments on the public encrypted network traffic dataset ISCX2016 demonstrate the effectiveness of the proposed model.

关键词： network traffic classification convolution neural network attention mechanism network management cyber security

Image-to-Image Style Transfer Based on the Ghost Module

在线全文

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2021年第9期68卷 4051-4067页

作者： Yan Jiang Xinrui Jia Liguo Zhang Ye Yuan Lei Chen Guisheng Yin College of Computer Science and Technology Harbin Engineering UniversityHarbin150001China Heilongjiang Hengxun Technology Co. Ltd.Harbin150001China College of Engineering and Information Technology Georgia Southern UniversityStatesboro30458GAUSA

The technology for image-to-image style transfer(a prevalent image processing task)has developed rapidly.The purpose of style transfer is to extract a texture from the source image domain and transfer it to the target image domain using a deep neural network.However,the existing methods typically have a large computational cost.To achieve efficient style transfer,we introduce a novel Ghost module into the GANILLA architecture to produce more feature maps from cheap operations.Then we utilize an attention mechanism to transform images with various styles.We optimize the original generative adversarial network(GAN)by using more efficient calculation methods for image-to-illustration translation.The experimental results show that our proposed method is similar to human vision and still maintains the quality of the image.Moreover,our proposed method overcomes the high computational cost and high computational resource consumption for style transfer.By comparing the results of subjective and objective evaluation indicators,our proposed method has shown superior performance over existing methods.

关键词： Style transfer generative adversarial networks ghost module attention mechanism human visual habits

Review of Text Classification Methods on Deep Learning

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Computers, Materials & Continua 2020年第6期63卷 1309-1321页

作者： Hongping Wu Yuling Liu Jingwen Wang College of Computer Science and Electronic Engineering Hunan UniversityChangsha410082China Department of Computer Science Elizabethtown CollegePA17022USA

Text classification has always been an increasingly crucial topic in natural language processing.Traditional text classification methods based on machine learning have many disadvantages such as dimension explosion,data sparsity,limited generalization ability and so on.Based on deep learning text classification,this paper presents an extensive study on the text classification models including Convolutional Neural Network-Based(CNN-Based),Recurrent Neural Network-Based(RNN-based),attention mechanisms-Based and so on.Many studies have proved that text classification methods based on deep learning outperform the traditional methods when processing large-scale and complex datasets.The main reasons are text classification methods based on deep learning can avoid cumbersome feature extraction process and have higher prediction accuracy for a large set of unstructured data.In this paper,we also summarize the shortcomings of traditional text classification methods and introduce the text classification process based on deep learning including text preprocessing,distributed representation of text,text classification model construction based on deep learning and performance evaluation.

关键词： Text classification deep learning distributed representation CNN RNN attention mechanism

Deep Sequential Model for Anchor Recommendation on Live Streaming Platforms

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Big Data Mining and Analytics 2021年第3期4卷 173-182页

作者： Shuai Zhang Hongyan Liu Jun He Sanpu Han Xiaoyong Du School of Information Renmin University of ChinaBeijing 100872China School of Economics and Management Tsinghua UniversityBeijing 100084China Beijing Mijing Hefeng Technology Co.Ltd. Beijing 100621China

Live streaming has grown rapidly in recent years, attracting increasingly more participation. As the number of online anchors is large, it is difficult for viewers to find the anchors they are interested in. Therefore, a personalized recommendation system is important for live streaming platforms. On live streaming platforms, the viewer’s and anchor’s preferences are dynamically changing over time. How to capture the user’s preference change is extensively studied in the literature, but how to model the viewer’s and anchor’s preference changes and how to learn their representations based on their preference matching are less studied. Taking these issues into consideration, in this paper, we propose a deep sequential model for live streaming recommendation. We develop a component named the multi-head related-unit in the model to capture the preference matching between anchor and viewer and extract related features for their representations. To evaluate the performance of our proposed model, we conduct experiments on real datasets, and the results show that our proposed model outperforms state-of-the-art recommendation models.

关键词： live streaming sequential recommendation attention mechanism deep learning