A Lightweight Convolutional Neural Network with Hierarchical Multi-Scale Feature Fusion for Image Classification

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.

Grid Side Distributed Energy Storage Cloud Group End Region Hierarchical Time-Sharing Configuration Algorithm Based onMulti-Scale and Multi Feature Convolution Neural Network

There is instability in the distributed energy storage cloud group end region on the power grid side.In order to avoid large-scale fluctuating charging and discharging in the power grid environment and make the capacitor components showa continuous and stable charging and discharging state,a hierarchical time-sharing configuration algorithm of distributed energy storage cloud group end region on the power grid side based on multi-scale and multi feature convolution neural network is proposed.Firstly,a voltage stability analysis model based onmulti-scale and multi feature convolution neural network is constructed,and the multi-scale and multi feature convolution neural network is optimized based on Self-OrganizingMaps(SOM)algorithm to analyze the voltage stability of the cloud group end region of distributed energy storage on the grid side under the framework of credibility.According to the optimal scheduling objectives and network size,the distributed robust optimal configuration control model is solved under the framework of coordinated optimal scheduling at multiple time scales;Finally,the time series characteristics of regional power grid load and distributed generation are analyzed.According to the regional hierarchical time-sharing configuration model of“cloud”,“group”and“end”layer,the grid side distributed energy storage cloud group end regional hierarchical time-sharing configuration algorithm is realized.The experimental results show that after applying this algorithm,the best grid side distributed energy storage configuration scheme can be determined,and the stability of grid side distributed energy storage cloud group end region layered timesharing configuration can be improved.

Nonparametric Statistical Feature Scaling Based Quadratic Regressive Convolution Deep Neural Network for Software Fault Prediction

The development of defect prediction plays a significant role in improving software quality. Such predictions are used to identify defective modules before the testing and to minimize the time and cost. The software with defects negatively impacts operational costs and finally affects customer satisfaction. Numerous approaches exist to predict software defects. However, the timely and accurate software bugs are the major challenging issues. To improve the timely and accurate software defect prediction, a novel technique called Nonparametric Statistical feature scaled QuAdratic regressive convolution Deep nEural Network (SQADEN) is introduced. The proposed SQADEN technique mainly includes two major processes namely metric or feature selection and classification. First, the SQADEN uses the nonparametric statistical Torgerson–Gower scaling technique for identifying the relevant software metrics by measuring the similarity using the dice coefficient. The feature selection process is used to minimize the time complexity of software fault prediction. With the selected metrics, software fault perdition with the help of the Quadratic Censored regressive convolution deep neural network-based classification. The deep learning classifier analyzes the training and testing samples using the contingency correlation coefficient. The softstep activation function is used to provide the final fault prediction results. To minimize the error, the Nelder–Mead method is applied to solve non-linear least-squares problems. Finally, accurate classification results with a minimum error are obtained at the output layer. Experimental evaluation is carried out with different quantitative metrics such as accuracy, precision, recall, F-measure, and time complexity. The analyzed results demonstrate the superior performance of our proposed SQADEN technique with maximum accuracy, sensitivity and specificity by 3%, 3%, 2% and 3% and minimum time and space by 13% and 15% when compared with the two state-of-the-art methods.

Multi⁃Scale Dilated Convolutional Neural Network for Hyperspectral Image Classification

Aiming at the problem of image information loss,dilated convolution is introduced and a novel multi⁃scale dilated convolutional neural network(MDCNN)is proposed.Dilated convolution can polymerize image multi⁃scale information without reducing the resolution.The first layer of the network used spectral convolutional step to reduce dimensionality.Then the multi⁃scale aggregation extracted multi⁃scale features through applying dilated convolution and shortcut connection.The extracted features which represent properties of data were fed through Softmax to predict the samples.MDCNN achieved the overall accuracy of 99.58% and 99.92% on two public datasets,Indian Pines and Pavia University.Compared with four other existing models,the results illustrate that MDCNN can extract better discriminative features and achieve higher classification performance.

Pedestrian attribute classification with multi-scale and multi-label convolutional neural networks

Pedestrian attribute classification from a pedestrian image captured in surveillance scenarios is challenging due to diverse clothing appearances,varied poses and different camera views. A multiscale and multi-label convolutional neural network( MSMLCNN) is proposed to predict multiple pedestrian attributes simultaneously. The pedestrian attribute classification problem is firstly transformed into a multi-label problem including multiple binary attributes needed to be classified. Then,the multi-label problem is solved by fully connecting all binary attributes to multi-scale features with logistic regression functions. Moreover,the multi-scale features are obtained by concatenating those featured maps produced from multiple pooling layers of the MSMLCNN at different scales. Extensive experiment results show that the proposed MSMLCNN outperforms state-of-the-art pedestrian attribute classification methods with a large margin.

Global forward-predicting dynamic routing for traffic concurrency space stereo multi-layer scale-free network

Many real communication networks, such as oceanic monitoring network and land environment observation network,can be described as space stereo multi-layer structure, and the traffic in these networks is concurrent. Understanding how traffic dynamics depend on these real communication networks and finding an effective routing strategy that can fit the circumstance of traffic concurrency and enhance the network performance are necessary. In this light, we propose a traffic model for space stereo multi-layer complex network and introduce two kinds of global forward-predicting dynamic routing strategies, global forward-predicting hybrid minimum queue（HMQ） routing strategy and global forward-predicting hybrid minimum degree and queue（HMDQ） routing strategy, for traffic concurrency space stereo multi-layer scale-free networks. By applying forward-predicting strategy, the proposed routing strategies achieve better performances in traffic concurrency space stereo multi-layer scale-free networks. Compared with the efficient routing strategy and global dynamic routing strategy, HMDQ and HMQ routing strategies can optimize the traffic distribution, alleviate the number of congested packets effectively and reach much higher network capacity.

Improving consensual performance of multi-agent systems in weighted scale-free networks

This paper studies consensus problems in weighted scale-free networks of asymmetrically coupled dynamical units, where the asymmetry in a given link is deter：mined by the relative degree of the involved nodes. It shows that the asymmetry of interactions has a great effect on the consensus. Especially, when the interactions are dominant from higher- to lower-degree nodes, both the convergence speed and the robustness to communication delay are enhanced.

融合Multi-scale CNN和Bi-LSTM的人脸表情识别研究

为了有效改善现有人脸表情识别模型中存在信息丢失严重、特征信息之间联系不密切的问题,提出一种融合多尺度卷积神经网络(Multi-scale CNN)和双向长短期记忆(Bi-LSTM)的模型。Bi-LSTM可以增强特征信息间的联系与信息的维持,在Multi-scale CNN中通过不同尺度的卷积核可以提取到更加丰富的特征信息,并通过加入批标准化(BN)层与特征融合处理,从而加快网络的收敛速度,有利于特征信息的重利用,再将两者提取到的特征信息进行融合,最后将改进的正则化方法应用到目标函数中,减小网络复杂度和过拟合。在JAFFE和FER-2013公开数据集上进行实验,准确率分别达到了95.455%和74.115%,由此证明所提算法的有效性和先进性。

基于改进Multi-Scale AlexNet的番茄叶部病害图像识别

番茄同种病害在不同发病阶段表征差异明显,不同病害又表现出一定的相似性,传统模式识别方法不能体现病害病理表征的动态变化,实用性较差。针对该问题,基于卷积神经网络提出一种适用于移动平台的多尺度识别模型,并基于此模型开发了面向农业生产人员的番茄叶部病害图像识别系统。该文详细描述了AlexNet的结构,分析其不足,结合番茄病害叶片图像特点,去除局部响应归一化层、修改全连接层、设置不同尺度卷积核提取特征,设计了基于AlexNet的多感受野识别模型,并基于Android实现了使用此模型的番茄叶部病害图像识别系统。Multi-ScaleAlexNet模型运行所耗内存为29.9MB,比原始AlexNet的内存需求652MB降低了95.4%,该模型对番茄叶部病害及每种病害早中晚期的平均识别准确率达到92.7%,基于此模型的Andriod端识别系统在田间的识别率达到89.2%,能够满足生产实践中移动平台下的病害图像识别需求。研究结果可为基于卷积神经网络的作物病害图像识别提供参考,为作物病害的自动化识别和工程化应用参考。

基于Multi-Light模型的奶牛个体识别研究

为了解决大规模智能化奶牛养殖场对奶牛个体识别存在模型大、识别速度慢的问题,试验构建了一种用于识别奶牛个体的多尺度轻量化卷积神经网络(Multi-Light)模型,将拍摄的奶牛图像经过标注后利用DeepLab V3模型从复杂背景中分割出单头奶牛图像;在Multi-Light模型中引入空洞卷积,保证该模型参数量不变的同时增强提取图像全局信息的能力;加入多尺度卷积模块增强该模型对不同尺度特征点的检测能力,在该模型中使用短路连接以保证特征不丢失,提升模型的识别精度;此外,利用通道注意力机制提高了该模型识别精度,同时使该模型具有更多的非线性;最后将分割得到的奶牛图像数据集输入Multi-Light模型进行训练。结果表明:Multi-Light模型对奶牛个体识别的精度达98.51%,高于其他经典模型对奶牛个体的识别率;与轻量级模型对比,Multi-Light模型的大小为5.86 MB,在具备高识别精度的前提下参数量较少。说明试验所搭建的Multi-Light模型克服了传统方法中需要对特征进行人为提取、提取特征方法不够鲁棒、识别模型参数量大及识别速度慢的缺点,为奶牛个体轻量化识别提供了参考。

Multi-layer Tectonic Model for Intraplate Deformation and Plastic-Flow Network in the Asian Continental Lithosphere

In a large area of the east—central Asian continent there is a unified seismic network system composed of two families of large—seismic belts that intersect conjugately. Such a seismic network in the middle—upper crust is actually a response to the plastic flow network in the lower lithosphere including the lower crust and lithospheric mantle. The existence of the unified plastic flow system confirms that the driving force for intraplate tectonic deformation results mainly from the compression of the India plate, while the long-range transmission of the force is carried out chiefly by means of plastic flow. The plastic flow network has a control over the intraplate tectonic deformation.

Color Correction for Multi-view Video Using Energy Minimization of View Networks

Systems using numerous cameras are emerging in many fields due to their ease of production and reduced cost, and one of the fields where they are expected to be used more actively in the near future is in image-based rendering （IBR）. Color correction between views is necessary to use multi-view systems in IBR to make audiences feel comfortable when views are switched or when a free viewpoint video is displayed. Color correction usually involves two steps： the first is to adjust camera parameters such as gain, brightness, and aperture before capture, and the second is to modify captured videos through image processing. This paper deals with the latter, which does not need a color pattern board. The proposed method uses scale invariant feature transform （SIFT） to detect correspondences, treats RGB channels independently, calculates lookup tables with an energy-minimization approach, and corrects captured video with these tables. The experimental results reveal that this approach works well.

Rotation,Translation and Scale Invariant Sign Word Recognition Using Deep Learning

Communication between people with disabilities and people who do not understand sign language is a growing social need and can be a tedious task.One of the main functions of sign language is to communicate with each other through hand gestures.Recognition of hand gestures has become an important challenge for the recognition of sign language.There are many existing models that can produce a good accuracy,but if the model test with rotated or translated images,they may face some difficulties to make good performance accuracy.To resolve these challenges of hand gesture recognition,we proposed a Rotation,Translation and Scale-invariant sign word recognition system using a convolu-tional neural network(CNN).We have followed three steps in our work:rotated,translated and scaled(RTS)version dataset generation,gesture segmentation,and sign word classification.Firstly,we have enlarged a benchmark dataset of 20 sign words by making different amounts of Rotation,Translation and Scale of the ori-ginal images to create the RTS version dataset.Then we have applied the gesture segmentation technique.The segmentation consists of three levels,i)Otsu Thresholding with YCbCr,ii)Morphological analysis:dilation through opening morphology and iii)Watershed algorithm.Finally,our designed CNN model has been trained to classify the hand gesture as well as the sign word.Our model has been evaluated using the twenty sign word dataset,five sign word dataset and the RTS version of these datasets.We achieved 99.30%accuracy from the twenty sign word dataset evaluation,99.10%accuracy from the RTS version of the twenty sign word evolution,100%accuracy from thefive sign word dataset evaluation,and 98.00%accuracy from the RTS versionfive sign word dataset evolution.Furthermore,the influence of our model exists in competitive results with state-of-the-art methods in sign word recognition.

多区域注意力的细粒度图像分类网络

目前细粒度图像分类的难点在于如何精准定位图像中高度可辨的局部区域以及其他辅助判别特征。提出一种多区域注意力的细粒度图像分类网络来解决这个问题。首先使用Inception-V3对图像特征进行提取,通过重复使用注意力擦除的方法使模型关注次要特征;然后通过背景去除以及上采样的方法获取图像更精准的局部图像,对提取到的局部特征进行位置统计,并以矩形框的方式获取图像整体,减少细节信息丢失;最后对局部与整体图像进行更加细致的学习。此外,设计联合损失函数,通过动态平衡难易样本和缩小类内差距的方法改善模型的识别效果。实验结果表明,该方法在公开的细粒度图像数据集CUB-200-2011、Stanford-Cars和FGVC-Aircraft上的准确率分别达到89.2%、94.8%、94.0%,相较于对比方法性能更优。

基于多尺度特征深度神经网络的不同产地山楂细粒度图像识别

中药是中医治疗疾病的主要途径,也是我国中医药事业传承与创新发展的物质基础,其真伪优劣也会直接影响中医临床的疗效,因此研究科学合理且高效的中药材质量检测方法符合当前行业热点.山楂作为中国著名的药食两用类药材,在烹饪和治疗中具有保护心血管、降低血压的作用,被广泛应用;但由于自然环境与栽培条件的不同,不同产地的山楂易被混淆从而对品质产生影响.尽管化学、生物鉴定的方法广泛而重要,但专业门槛高,耗时较长;且传统图像处理方法容易受外在环境因素干扰,可靠性差.因此亟待研究快速准确的方法以实现山楂产地的精准鉴别;受CoAtNet与Swin-Transformer网络启发,本文结合MBConv模块中深度可分离卷积网络对局部信息建模的特点与Swin Transformer模块多层次结构可弥补网络非局部性损失的特性,提出一种多尺度特征的混合神经网络模型,通过获取图像不同层级特征,将获取的形状、颜色与纹理等浅层特征作为先验知识与高层级语义信息进行特征融合,研究了一种快速有效的识别方法以实现对不同产地山楂的有效鉴别;此外,本文提出一种新的局部空间注意力机制,通过形成通道注意力模块联合空间注意力模块的新结构,实现对图像细粒度特征的聚焦与学习.实验结果表明,本文所提出的方法有最高的鉴别准确率为89.306%,优于其他基线模型.实践证明,本文的研究提高中药材鉴别的科技水平,拓宽传统中医药的研究思路.

基于卷积神经网络的预制叠合板多目标智能化检测方法

在生产过程中,预制构件尺寸不合格问题将导致其在施工现场无法顺利安装,从而影响工期。为推进预制构件智能化生产的进程,以预制叠合板为例,基于卷积神经网络研究生产过程中的智能检测方法,在生产流水线上设计并安装图像采集系统,建立预制叠合板尺寸检测数据集。通过YOLOv5算法实现对混凝土底板、预埋PVC线盒及外伸钢筋的识别,并以固定磁盒作为基准参照物进行尺寸检测误差分析,实现混凝土底板尺寸、预埋PVC线盒坐标的检测,在降低训练数据集参数规模的工况下保持较高的识别精度。结果表明:该方法可以有效检测预制叠合板的底板数量和尺寸、预埋PVC线盒数量和坐标,并实现弯折方向不合格的外伸钢筋检测,并能降低人工成本,提高检测精度,加快检测速度,提高预制叠合板的出厂质量。

基于多尺度特征信息融合的时间序列异常检测

目前大多数的时间序列都缺少相应的异常标签,且现有基于重构的异常检测算法不能很好地捕获到多维数据间复杂的潜在相关性和时间依赖性,为了构建特征丰富的时间序列,提出一种多尺度特征信息融合的异常检测模型。该模型首先通过卷积神经网络对滑动窗口内的不同序列进行特征卷积来获取不同尺度下的局部上下文信息。然后,利用Transformer中的位置编码对卷积后的时间序列窗口进行位置嵌入,增强滑动窗口中每一个时间序列和邻近序列之间的位置联系,并引入时间注意力获取数据在时间维度上的自相关性,并进一步通过多头自注意力自适应地为窗口内不同时间序列分配不同的权重。最后,对反卷积过程中上采样得到的窗口数据与不同尺度下得到的局部特征和时间上下文信息进行逐步融合,从而准确重构原始时间序列,并将重构误差作为最终的异常得分进行异常判定。实验结果表明,所构建模型在SWaT和SMD数据集上与基线模型相比F1分数均有所提升。在数据维度高且均衡性较差的WADI数据集上与GDN模型相比F1分数降低了1.66%。

数据驱动的半无限介质裂纹识别模型研究

缺陷识别是结构健康监测的重要研究内容,对评估工程结构的安全性具有重要的指导意义,然而,准确确定结构缺陷的尺寸十分困难.论文提出了一种创新的数据驱动算法,将比例边界有限元法(scaled boundary finite element methods,SBFEM)与自编码器(autoencoder,AE)、因果膨胀卷积神经网络(causal dilated convolutional neural network,CDCNN)相结合用于半无限介质中的裂纹识别.在该模型中,SBFEM用于模拟波在含不同裂纹状缺陷半无限介质中的传播过程,对于不同的裂纹状缺陷,仅需改变裂纹尖端的比例中心和裂纹开口处节点的位置,避免了复杂的重网格过程,可高效地生成足够的训练数据.模拟波在半无限介质中传播时,建立了基于瑞利阻尼的吸收边界模型,避免了对结构全域模型进行计算.搭建了CDCNN,确保了时序数据的有序性,并获得更大的感受野而不增加神经网络的复杂性,可捕捉更多的历史信息,AE具有较强的非线性特征提取能力,可将高维的原始输入特征向量空间映射到低维潜在特征向量空间,以获得低维潜在特征用于网络模型训练,有效提升了网络模型的学习效率.数值算例表明:提出的模型能够高效且准确地识别半无限介质中裂纹的量化信息,且AE-CDCNN模型的识别效率较单CDCNN模型提高了约2.7倍.

基于CycleGAN-IA方法和M-ConvNext网络的苹果叶片病害图像识别

针对苹果叶片病害图像识别存在数据集获取困难、样本不足、识别准确率低等问题,提出基于多尺度特征提取的病害识别网络(Multi-scale feature extraction ConvNext,M-ConvNext)模型。采用一种结合改进的循环一致性生成对抗网络与仿射变换的数据增强方法(Improved CycleGAN and affine transformation,CycleGAN-IA),首先,使用较小感受野的卷积核和残差注意力模块优化CycleGAN网络结构,使用二值交叉熵损失函数代替CycleGAN网络的均方差损失函数,以此生成高质量样本图像,提高样本特征复杂度;然后,对生成图像进行仿射变换,提高数据样本的空间复杂度,该方法解决了数据样本不足的问题,用于辅助后续的病害识别模型。其次,构建M-ConvNext网络,该网络设计G-RFB模块获取并融合各个尺度的特征信息,GELU激活函数增强网络的特征表达能力,提高苹果叶片病害图像识别准确率。最后,实验结果表明,CycleGAN-IA数据增强方法可以对数据集起到良好的扩充作用,在常用网络上验证,增强后的数据集可以有效提高苹果叶片病害图像识别准确率;通过消融实验可得,M-ConvNex识别准确率可达99.18%,较原ConvNext网络准确率提高0.41个百分点,较ResNet50、MobileNetV3和EfficientNetV2网络分别提高3.78、7.35、4.07个百分点,为后续农作物病害识别提供了新思路。

一种道路裂缝检测的变尺度VS-UNet模型

为解决目前现有的图像分割算法存在检测精度低、对裂缝检测缺乏针对性等问题,采用多尺度特征融合方法,提出一种扩展LG Block模块Extend-LG Block,其由多个并行不同膨胀率的空洞卷积组成.通过参数可调节分支数量和空洞卷积膨胀率,从而改变其感受野大小,进而提取和融合不同尺度的裂缝特征.对比在深层使用多尺度特征融合模块的网络以及使用固定尺度结构进行多尺度特征融合的网络的优劣,提出一种变尺度结构的UNet模型VS-UNet,使用多个不同参数的Extend-LG Block替换UNet网络中的基本卷积块.该结构在网络浅层进行多尺度特征融合,多尺度特征融合模块提取的尺度随网络层加深逐渐减少.此结构在加强图像的细节特征提取能力的同时保持原有的抽象特征提取能力,还可避免网络参数的增加.在DeepCrack数据集以及CFD数据集上进行实验验证,结果表明,相较于其他两种结构和方法,提出的变尺度结构的网络在有更高检测精度的同时,在可视化实验对比上对各种大小的裂缝有更好的分割效果.最后与其他图像分割算法进行对比,各项指标与UNet相比均有一定程度提升,证明了网络改进的有效性.研究结果可为进一步提升道路裂缝检测效果提供参考.