Unsupervised multi-modal image translation based on the squeeze-and-excitation mechanism and feature attention module

The unsupervised multi-modal image translation is an emerging domain of computer vision whose goal is to transform an image from the source domain into many diverse styles in the target domain.However,the multi-generator mechanism is employed among the advanced approaches available to model different domain mappings,which results in inefficient training of neural networks and pattern collapse,leading to inefficient generation of image diversity.To address this issue,this paper introduces a multi-modal unsupervised image translation framework that uses a generator to perform multi-modal image translation.Specifically,firstly,the domain code is introduced in this paper to explicitly control the different generation tasks.Secondly,this paper brings in the squeeze-and-excitation(SE)mechanism and feature attention(FA)module.Finally,the model integrates multiple optimization objectives to ensure efficient multi-modal translation.This paper performs qualitative and quantitative experiments on multiple non-paired benchmark image translation datasets while demonstrating the benefits of the proposed method over existing technologies.Overall,experimental results have shown that the proposed method is versatile and scalable.

Two-Layer Attention Feature Pyramid Network for Small Object Detection

Effective small object detection is crucial in various applications including urban intelligent transportation and pedestrian detection.However,small objects are difficult to detect accurately because they contain less information.Many current methods,particularly those based on Feature Pyramid Network(FPN),address this challenge by leveraging multi-scale feature fusion.However,existing FPN-based methods often suffer from inadequate feature fusion due to varying resolutions across different layers,leading to suboptimal small object detection.To address this problem,we propose the Two-layerAttention Feature Pyramid Network(TA-FPN),featuring two key modules:the Two-layer Attention Module(TAM)and the Small Object Detail Enhancement Module(SODEM).TAM uses the attention module to make the network more focused on the semantic information of the object and fuse it to the lower layer,so that each layer contains similar semantic information,to alleviate the problem of small object information being submerged due to semantic gaps between different layers.At the same time,SODEM is introduced to strengthen the local features of the object,suppress background noise,enhance the information details of the small object,and fuse the enhanced features to other feature layers to ensure that each layer is rich in small object information,to improve small object detection accuracy.Our extensive experiments on challenging datasets such as Microsoft Common Objects inContext(MSCOCO)and Pattern Analysis Statistical Modelling and Computational Learning,Visual Object Classes(PASCAL VOC)demonstrate the validity of the proposedmethod.Experimental results show a significant improvement in small object detection accuracy compared to state-of-theart detectors.

An attention-based prototypical network for forest fire smoke few-shot detection

Existing almost deep learning methods rely on a large amount of annotated data, so they are inappropriate for forest fire smoke detection with limited data. In this paper, a novel hybrid attention-based few-shot learning method, named Attention-Based Prototypical Network, is proposed for forest fire smoke detection. Specifically, feature extraction network, which consists of convolutional block attention module, could extract high-level and discriminative features and further decrease the false alarm rate resulting from suspected smoke areas. Moreover, we design a metalearning module to alleviate the overfitting issue caused by limited smoke images, and the meta-learning network enables achieving effective detection via comparing the distance between the class prototype of support images and the features of query images. A series of experiments on forest fire smoke datasets and miniImageNet dataset testify that the proposed method is superior to state-of-the-art few-shot learning approaches.

Fusion of Convolutional Self-Attention and Cross-Dimensional Feature Transformationfor Human Posture Estimation

Human posture estimation is a prominent research topic in the fields of human-com-puter interaction,motion recognition,and other intelligent applications.However,achieving highaccuracy in key point localization,which is crucial for intelligent applications,contradicts the lowdetection accuracy of human posture detection models in practical scenarios.To address this issue,a human pose estimation network called AT-HRNet has been proposed,which combines convolu-tional self-attention and cross-dimensional feature transformation.AT-HRNet captures significantfeature information from various regions in an adaptive manner,aggregating them through convolu-tional operations within the local receptive domain.The residual structures TripNeck and Trip-Block of the high-resolution network are designed to further refine the key point locations,wherethe attention weight is adjusted by a cross-dimensional interaction to obtain more features.To vali-date the effectiveness of this network,AT-HRNet was evaluated using the COCO2017 dataset.Theresults show that AT-HRNet outperforms HRNet by improving 3.2%in mAP,4.0%in AP75,and3.9%in AP^(M).This suggests that AT-HRNet can offer more beneficial solutions for human posture estimation.

Single Image Deraining Using Dual Branch Network Based on Attention Mechanism for IoT

Extracting useful details from images is essential for the Internet of Things project.However,in real life,various external environments,such as badweather conditions,will cause the occlusion of key target information and image distortion,resulting in difficulties and obstacles to the extraction of key information,affecting the judgment of the real situation in the process of the Internet of Things,and causing system decision-making errors and accidents.In this paper,we mainly solve the problem of rain on the image occlusion,remove the rain grain in the image,and get a clear image without rain.Therefore,the single image deraining algorithm is studied,and a dual-branch network structure based on the attention module and convolutional neural network(CNN)module is proposed to accomplish the task of rain removal.In order to complete the rain removal of a single image with high quality,we apply the spatial attention module,channel attention module and CNN module to the network structure,and build the network using the coder-decoder structure.In the experiment,with the structural similarity(SSIM)and the peak signal-to-noise ratio(PSNR)as evaluation indexes,the training and testing results on the rain removal dataset show that the proposed structure has a good effect on the single image deraining task.

Multi-Scale Attention-Based Deep Neural Network for Brain Disease Diagnosis

Whole brain functional connectivity(FC)patterns obtained from resting-state functional magnetic resonance imaging(rs-fMRI)have been widely used in the diagnosis of brain disorders such as autism spectrum disorder(ASD).Recently,an increasing number of studies have focused on employing deep learning techniques to analyze FC patterns for brain disease classification.However,the high dimensionality of the FC features and the interpretation of deep learning results are issues that need to be addressed in the FC-based brain disease classification.In this paper,we proposed a multi-scale attention-based deep neural network(MSA-DNN)model to classify FC patterns for the ASD diagnosis.The model was implemented by adding a flexible multi-scale attention(MSA)module to the auto-encoder based backbone DNN,which can extract multi-scale features of the FC patterns and change the level of attention for different FCs by continuous learning.Our model will reinforce the weights of important FC features while suppress the unimportant FCs to ensure the sparsity of the model weights and enhance the model interpretability.We performed systematic experiments on the large multi-sites ASD dataset with both ten-fold and leaveone-site-out cross-validations.Results showed that our model outperformed classical methods in brain disease classification and revealed robust intersite prediction performance.We also localized important FC features and brain regions associated with ASD classification.Overall,our study further promotes the biomarker detection and computer-aided classification for ASD diagnosis,and the proposed MSA module is flexible and easy to implement in other classification networks.

基于时空规律的PCA-LSTM-Attention空气质量预测研究

空气质量指数(AQI)是考量空气质量好坏的综合指标,由于各地区空气受风向影响不断流动,使传统预测模型难以从时间单一角度进行建模。针对此问题提出一种利用主成分分析(PCA)降维思想,考虑不同地区时空规律的模型。通过收集目标城市和周边几个城市的所需数据,使用PCA求解所有城市的综合空气得分作为空间信息,再输入LSTM提取时间规律,最后通过注意力模块输出AQI预测。通过对沧州、唐山、廊坊、保定和天津的大气污染物和气象数据的分析,证明该算法比只考虑时间因素的LSTM模型、RNN模型和ARIMA(1,1,1)模型精度更高,可以有助于提高AQI预测精度。

Attention Res-Unet:一种高效阴影检测算法

图像中阴影像素的存在会导致图像内容的不确定性,对计算机视觉任务有害,因此常将阴影检测作为计算机视觉算法的预处理步骤.提出全新的阴影检测网络结构,通过结合输入图像中包含的语义信息和像素之间的关联,提升网络性能.使用预训练后的深层网络ResNeXt101作为特征提取前端,提取图像的语义信息,并结合U-net的设计思路,搭建网络结构,完成特征层的上采样过程.在输出层之前使用非局部操作,为每一个像素提供全局信息,建立像素与像素之间的联系.设计注意力生成模块和注意力融合模块,进一步提高检测准确率.分别在SBU、UCF这2个阴影检测数据集上进行验证,实验结果表明,所提方法的目视效果及客观指标皆优于此前最优方法所得结果,在2个数据集上的平均检测错误率分别降低14.4%和14.9%.

基于STFT和CNN-Attention的配电终端采集模块故障诊断研究

针对复杂工况运行环境下配电终端采集模块故障类型难以识别的问题,提出一种基于短时傅里叶变换(STFT)、卷积神经网络和注意力机制(CNN-Attention)的配电终端采集模块故障诊断方法。首先,分析配电终端采集模块不同故障类型会产生的对应故障数据,建立故障数据集。然后,基于STFT提取故障数据的故障时频特征以形成时频图,采用CNN-Attention模型对时频图进行故障诊断与匹配。算例分析表明,CNN-Attention的故障检测准确率为97.31%,相较于CNN和极限学习机(ELM)模型,故障诊断准确率分别提升了1.22%和4.4%。Attention机制能够有效解决CNN在特征提取时产生的冗余信息导致模型训练慢、难以收敛的问题。该研究实现了配电终端采集模块具体故障类型的准确识别,能为后续配电终端的运维提供参考。

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

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.

Social Robot Detection Method with Improved Graph Neural Networks

Social robot accounts controlled by artificial intelligence or humans are active in social networks,bringing negative impacts to network security and social life.Existing social robot detection methods based on graph neural networks suffer from the problem of many social network nodes and complex relationships,which makes it difficult to accurately describe the difference between the topological relations of nodes,resulting in low detection accuracy of social robots.This paper proposes a social robot detection method with the use of an improved neural network.First,social relationship subgraphs are constructed by leveraging the user’s social network to disentangle intricate social relationships effectively.Then,a linear modulated graph attention residual network model is devised to extract the node and network topology features of the social relation subgraph,thereby generating comprehensive social relation subgraph features,and the feature-wise linear modulation module of the model can better learn the differences between the nodes.Next,user text content and behavioral gene sequences are extracted to construct social behavioral features combined with the social relationship subgraph features.Finally,social robots can be more accurately identified by combining user behavioral and relationship features.By carrying out experimental studies based on the publicly available datasets TwiBot-20 and Cresci-15,the suggested method’s detection accuracies can achieve 86.73%and 97.86%,respectively.Compared with the existing mainstream approaches,the accuracy of the proposed method is 2.2%and 1.35%higher on the two datasets.The results show that the method proposed in this paper can effectively detect social robots and maintain a healthy ecological environment of social networks.

基于解耦注意力与幻影卷积的轻量级人体姿态估计

随着轻量级网络的发展,人体姿态估计任务得以在计算资源有限的设备上执行,然而,提升精度变得更具有挑战性。这些挑战主要源于网络复杂度与计算资源的矛盾,导致模型在简化时牺牲了表示能力。针对上述问题,提出一种基于解耦注意力和幻影卷积的轻量级人体姿态估计网络(DGLNet)。具体来说,DGLNet以小型高分辨率网络(Small HRNet)模型为基础架构,通过引入解耦注意力机制构建DFDbottleneck模块;采用shuffleblock的结构对基础模块进行重新设计,即用轻量级幻影卷积替代计算量大的点卷积,并利用解耦注意力机制增强模块性能,从而构建DGBblock模块;此外,用幻影卷积和解耦注意力重新构建的深度可分离卷积模块来替代原过渡层模块,从而构建GSCtransition模块,进一步减少计算量并增强特征交互性和提高性能。在COCO验证集上的实验结果显示,DGLNet优于轻量级高分辨率网络(Lite-HRNet),在计算量和参数量不增加的情况下,最高精度达到了71.9%;与常见的轻量级姿态估计网络MobileNetV2和ShuffleNetV2相比,DGLNet在仅使用21.2%和25.0%的计算量情况下分别实现了4.6和8.3个百分点的精度提升;在AP~(50)的评价标准上,DGLNet超过了大型高分辨率网络(HRNet)的同时计算量和参数量远小于HRNet。

基于柔性残差神经网络的滚动轴承智能故障诊断方法

滚动轴承作为旋转机械的重要组成部分,其正常运行直接影响机器的使用寿命和运行状态.为了提高滚动轴承故障诊断的准确性,本文提出一种基于动态减法平均优化器(DSABO)和平行注意力模块(PAM)的柔性残差神经网络(FResNet),用于滚动轴承故障诊断.具体而言,首先设计一种基于卷积神经网络的柔性残差模块来构建FResNet.该模块允许在DSABO迭代时更改卷积层数、卷积核数和跳跃连接数,从而增强网络故障特征提取能力并减少网络退化.其次,设计具有卷积层的PAM来融合通道注意力和空间注意力输出权重,通过与滚动轴承运行数据结合,实现数据特征增强.于是,DSABO、PAM和FResNet的集成形成了一个有效的滚动轴承故障诊断模型,命名为DSABO-PAM-FResNet.最后,利用美国凯斯西储大学滚动轴承故障数据集验证所提DSABO-PAM-FResNet模型的可行性和有效性.结果显示,在信噪比为–6 dB环境下所提模型对滚动轴承故障诊断的准确率为97.18%,证明所提模型具有较好的抗噪能力;在0.75 kW、1.5 kW和2.25 kW不同负载条件下,所提模型对滚动轴承故障诊断的平均准确率为98.2%,证明所提模型具有良好的变工况诊断适应能力.与其他智能故障诊断方法的对比结果表明,所提DSABO-PAM-FResNet模型的诊断精度更高,为滚动轴承故障诊断提供了一种新的有效智能方法.

Bidirectional position attention lightweight network for massive MIMO CSI feedback

In frequency division duplex(FDD)massive multiple-input multiple-output(MIMO)systems,a bidirectional positional attention network(BPANet)was proposed to address the high computational complexity and low accuracy of existing deep learning-based channel state information(CSI)feedback methods.Specifically,a bidirectional position attention module(BPAM)was designed in the BPANet to improve the network performance.The BPAM captures the distribution characteristics of the CSI matrix by integrating channel and spatial dimension information,thereby enhancing the feature representation of the CSI matrix.Furthermore,channel attention is decomposed into two one-dimensional(1D)feature encoding processes effectively reducing computational costs.Simulation results demonstrate that,compared with the existing representative method complex input lightweight neural network(CLNet),BPANet reduces computational complexity by an average of 19.4%and improves accuracy by an average of 7.1%.Additionally,it performs better in terms of running time delay and cosine similarity.

Disease Recognition of Apple Leaf Using Lightweight Multi-Scale Network with ECANet

To solve the problem of difficulty in identifying apple diseases in the natural environment and the low application rate of deep learning recognition networks,a lightweight ResNet(LW-ResNet)model for apple disease recognition is proposed.Based on the deep residual network(ResNet18),the multi-scale feature extraction layer is constructed by group convolution to realize the compression model and improve the extraction ability of different sizes of lesion features.By improving the identity mapping structure to reduce information loss.By introducing the efficient channel attention module(ECANet)to suppress noise from a complex background.The experimental results show that the average precision,recall and F1-score of the LW-ResNet on the test set are 97.80%,97.92%and 97.85%,respectively.The parameter memory is 2.32 MB,which is 94%less than that of ResNet18.Compared with the classic lightweight networks SqueezeNet and MobileNetV2,LW-ResNet has obvious advantages in recognition performance,speed,parameter memory requirement and time complexity.The proposed model has the advantages of low computational cost,low storage cost,strong real-time performance,high identification accuracy,and strong practicability,which can meet the needs of real-time identification task of apple leaf disease on resource-constrained devices.

Siamese Dense Pixel-Level Fusion Network for Real-Time UAV Tracking

Onboard visual object tracking in unmanned aerial vehicles(UAVs)has attractedmuch interest due to its versatility.Meanwhile,due to high precision,Siamese networks are becoming hot spots in visual object tracking.However,most Siamese trackers fail to balance the tracking accuracy and time within onboard limited computational resources of UAVs.To meet the tracking precision and real-time requirements,this paper proposes a Siamese dense pixel-level network for UAV object tracking named SiamDPL.Specifically,the Siamese network extracts features of the search region and the template region through a parameter-shared backbone network,then performs correlationmatching to obtain the candidate regionwith high similarity.To improve the matching effect of template and search features,this paper designs a dense pixel-level feature fusion module to enhance the matching ability by pixel-wise correlation and enrich the feature diversity by dense connection.An attention module composed of self-attention and channel attention is introduced to learn global context information and selectively emphasize the target feature region in the spatial and channel dimensions.In addition,a target localization module is designed to improve target location accuracy.Compared with other advanced trackers,experiments on two public benchmarks,which are UAV123@10fps and UAV20L fromthe unmanned air vehicle123(UAV123)dataset,show that SiamDPL can achieve superior performance and low complexity with a running speed of 100.1 fps on NVIDIA TITAN RTX.

基于YOLO v5的农田杂草识别轻量化方法研究

针对已有杂草识别模型对复杂农田环境下多种目标杂草的识别率低、模型内存占用量大、参数多、识别速度慢等问题,提出了基于YOLO v5的轻量化杂草识别方法。利用带色彩恢复的多尺度视网膜(Multi-scale retinex with color restoration, MSRCR)增强算法对部分图像数据进行预处理,提高边缘细节模糊的图像清晰度,降低图像中的阴影干扰。使用轻量级网络PP-LCNet重置了识别模型中的特征提取网络,减少模型参数量。采用Ghost卷积模块轻量化特征融合网络,进一步降低计算量。为了弥补轻量化造成的模型性能损耗,在特征融合网络末端添加基于标准化的注意力模块(Normalization-based attention module, NAM),增强模型对杂草和玉米幼苗的特征提取能力。此外,通过优化主干网络注意力机制的激活函数来提高模型的非线性拟合能力。在自建数据集上进行实验,实验结果显示,与当前主流目标检测算法YOLO v5s以及成熟的轻量化目标检测算法MobileNet v3-YOLO v5s、ShuffleNet v2-YOLO v5s比较,轻量化后杂草识别模型内存占用量为6.23 MB,分别缩小54.5%、12%和18%;平均精度均值(Mean average precision, mAP)为97.8%,分别提高1.3、5.1、4.4个百分点。单幅图像检测时间为118.1 ms,达到了轻量化要求。在保持较高模型识别精度的同时大幅降低了模型复杂度,可为采用资源有限的移动端设备进行农田杂草识别提供技术支持。

基于改进YOLOv5s的轻量级绝缘子缺失检测

针对现有绝缘子缺失检测模型计算复杂度高和小目标难以检测等问题,提出一种基于改进的YOLOv5s轻量级检测模型。首先,移除主干网络中的C3模块来减少模型的参数量。其次,在多尺度特征融合网络中引入卷积块注意力机制来提高复杂背景下模型的特征提取能力。同时,采用加权双向特征金字塔网络结构对特征进行双向跨尺度加权融合,提升网络在遮挡物、相似目标干扰下目标的检测性能。最后,选用SIoU损失函数提升网络的收敛速度和检测精度。实验结果表明,所提模型的平均精准率为96.8%,浮点运算数为2.8 GFLOPS,而原始YOLOv5s在保证97.4%的平均精准率下的浮点运算数为16.3 GFLOPS。相较于原始模型,所提模型对小目标、遮挡目标以及模糊等场景有着较强的鲁棒性,且在保证近似检测精度的同时极大减少了计算量。

基于DenseNet和卷积注意力模块的高精度手势识别

非接触的手势识别是一种新型人机交互方式,在增强现实(AR)/虚拟现实(VR)、智能家居、智能医疗等方面有着广阔的应用前景,近年来成为一个研究热点。由于需要利用毫米波雷达进行更精确的微动手势识别,该文提出一种新型的基于MIMO毫米波雷达的微动手势识别方法。采用4片AWR1243雷达板级联而成的毫米波级联(MMWCAS)雷达采集手势回波,对手势回波进行时频分析,基于距离-多普勒(RD)图和3D点云检测出人手目标。通过数据预处理,提取手势目标的距离-时间谱图(RTM)、多普勒-时间谱图(DTM)、方位角-时间谱图(ATM)和俯仰角-时间谱图(ETM),更加全面地表征手势的运动特征,并形成混合特征谱图(FTM),对12种微动手势进行识别。设计了基于DenseNet和卷积注意力模块的手势识别网络,将混合特征谱图作为网络的输入,创新性地融合了卷积注意力模块(CBAM),实验表明,识别准确率达到99.03%,且该网络将注意力放在手势动作的前半段,实现了高精度的手势识别。

融合Inception V1-CBAM-CNN的轴承剩余寿命预测模型

针对现有的滚动轴承剩余寿命(RUL)预测方法精度低、轴承健康指标(HI)构建困难等问题,提出了一种基于卷积神经网络(CNN)并融合Inception V1模块和卷积注意力机制模块(CBAM)的滚动轴承RUL预测模型。首先,在CNN中添加了CBAM机制,并进行了加权处理,在通道和空间维度对重要特征进行了强化,对次要特征进行了抑制,通过添加改进的InceptionV1模块,提高了CNN通道间信息交互水平,全面提取了退化特征;然后,进行了网络优化,采用全局最大池化(GMP)方法对模型进行了简化,采用Dropout和批量归一化(BN)方法,避免了过拟合,提高了精度,且克服了训练时出现的梯度消失问题;最后,对数据进行了处理,将降噪后的信号重组为三维张量,将其作为HI,构建了退化标签,引入了评价指标,采用PHM2012轴承数据集进行了实验验证,在3种工况下将其与深度神经网络(DNN)、CNN方法、结合注意力机制的残差网络方法(ResNet)进行了对比。研究结果表明:该方法在变负载条件下的平均RMSE为0.033,较其他方法的RMSE值分别降低了86%、78%和69%,在预测精度和泛化能力方面具有明显优势。