Dynamic Multi-Layer Perceptron for Fetal Health Classification Using Cardiotocography Data

Fetal health care is vital in ensuring the health of pregnant women and the fetus.Regular check-ups need to be taken by the mother to determine the status of the fetus’growth and identify any potential problems.To know the status of the fetus,doctors monitor blood reports,Ultrasounds,cardiotocography(CTG)data,etc.Still,in this research,we have considered CTG data,which provides information on heart rate and uterine contractions during pregnancy.Several researchers have proposed various methods for classifying the status of fetus growth.Manual processing of CTG data is time-consuming and unreliable.So,automated tools should be used to classify fetal health.This study proposes a novel neural network-based architecture,the Dynamic Multi-Layer Perceptron model,evaluated from a single layer to several layers to classify fetal health.Various strategies were applied,including pre-processing data using techniques like Balancing,Scaling,Normalization hyperparameter tuning,batch normalization,early stopping,etc.,to enhance the model’s performance.A comparative analysis of the proposed method is done against the traditional machine learning models to showcase its accuracy(97%).An ablation study without any pre-processing techniques is also illustrated.This study easily provides valuable interpretations for healthcare professionals in the decision-making process.

A Power Data Anomaly Detection Model Based on Deep Learning with Adaptive Feature Fusion

With the popularisation of intelligent power,power devices have different shapes,numbers and specifications.This means that the power data has distributional variability,the model learning process cannot achieve sufficient extraction of data features,which seriously affects the accuracy and performance of anomaly detection.Therefore,this paper proposes a deep learning-based anomaly detection model for power data,which integrates a data alignment enhancement technique based on random sampling and an adaptive feature fusion method leveraging dimension reduction.Aiming at the distribution variability of power data,this paper developed a sliding window-based data adjustment method for this model,which solves the problem of high-dimensional feature noise and low-dimensional missing data.To address the problem of insufficient feature fusion,an adaptive feature fusion method based on feature dimension reduction and dictionary learning is proposed to improve the anomaly data detection accuracy of the model.In order to verify the effectiveness of the proposed method,we conducted effectiveness comparisons through elimination experiments.The experimental results show that compared with the traditional anomaly detection methods,the method proposed in this paper not only has an advantage in model accuracy,but also reduces the amount of parameter calculation of the model in the process of feature matching and improves the detection speed.

Digital modulation classification using multi-layer perceptron and time-frequency features

Considering that real communication signals corrupted by noise are generally nonstationary, and timefrequency distributions are especially suitable for the analysis of nonstationary signals, time-frequency distributions are introduced for the modulation classification of communication signals： The extracted time-frequency features have good classification information, and they are insensitive to signal to noise ratio （SNR） variation. According to good classification by the correct rate of a neural network classifier, a multilayer perceptron （MLP） classifier with better generalization, as well as, addition of time-frequency features set for classifying six different modulation types has been proposed. Computer simulations show that the MLP classifier outperforms the decision-theoretic classifier at low SNRs, and the classification experiments for real MPSK signals verify engineering significance of the MLP classifier.

Impact of Portable Executable Header Features on Malware Detection Accuracy

One aspect of cybersecurity,incorporates the study of Portable Executables(PE)files maleficence.Artificial Intelligence(AI)can be employed in such studies,since AI has the ability to discriminate benign from malicious files.In this study,an exclusive set of 29 features was collected from trusted implementations,this set was used as a baseline to analyze the presented work in this research.A Decision Tree(DT)and Neural Network Multi-Layer Perceptron(NN-MLPC)algorithms were utilized during this work.Both algorithms were chosen after testing a few diverse procedures.This work implements a method of subgrouping features to answer questions such as,which feature has a positive impact on accuracy when added?Is it possible to determine a reliable feature set to distinguish a malicious PE file from a benign one?when combining features,would it have any effect on malware detection accuracy in a PE file?Results obtained using the proposed method were improved and carried few observations.Generally,the obtained results had practical and numerical parts,for the practical part,the number of features and which features included are the main factors impacting the calculated accuracy,also,the combination of features is as crucial in these calculations.Numerical results included,finding accuracies with enhanced values,for example,NN_MLPC attained 0.979 and 0.98;for DT an accuracy of 0.9825 and 0.986 was attained.

Feature Extraction of Kernel Regress Reconstruction for Fault Diagnosis Based on Self-organizing Manifold Learning

The feature space extracted from vibration signals with various faults is often nonlinear and of high dimension.Currently,nonlinear dimensionality reduction methods are available for extracting low-dimensional embeddings,such as manifold learning.However,these methods are all based on manual intervention,which have some shortages in stability,and suppressing the disturbance noise.To extract features automatically,a manifold learning method with self-organization mapping is introduced for the first time.Under the non-uniform sample distribution reconstructed by the phase space,the expectation maximization(EM) iteration algorithm is used to divide the local neighborhoods adaptively without manual intervention.After that,the local tangent space alignment(LTSA) algorithm is adopted to compress the high-dimensional phase space into a more truthful low-dimensional representation.Finally,the signal is reconstructed by the kernel regression.Several typical states include the Lorenz system,engine fault with piston pin defect,and bearing fault with outer-race defect are analyzed.Compared with the LTSA and continuous wavelet transform,the results show that the background noise can be fully restrained and the entire periodic repetition of impact components is well separated and identified.A new way to automatically and precisely extract the impulsive components from mechanical signals is proposed.

DM-L Based Feature Extraction and Classifier Ensemble for Object Recognition

Deep Learning is a powerful technique that is widely applied to Image Recognition and Natural Language Processing tasks amongst many other tasks. In this work, we propose an efficient technique to utilize pre-trained Convolutional Neural Network (CNN) architectures to extract powerful features from images for object recognition purposes. We have built on the existing concept of extending the learning from pre-trained CNNs to new databases through activations by proposing to consider multiple deep layers. We have exploited the progressive learning that happens at the various intermediate layers of the CNNs to construct Deep Multi-Layer (DM-L) based Feature Extraction vectors to achieve excellent object recognition performance. Two popular pre-trained CNN architecture models i.e. the VGG_16 and VGG_19 have been used in this work to extract the feature sets from 3 deep fully connected multiple layers namely “fc6”, “fc7” and “fc8” from inside the models for object recognition purposes. Using the Principal Component Analysis (PCA) technique, the Dimensionality of the DM-L feature vectors has been reduced to form powerful feature vectors that have been fed to an external Classifier Ensemble for classification instead of the Softmax based classification layers of the two original pre-trained CNN models. The proposed DM-L technique has been applied to the Benchmark Caltech-101 object recognition database. Conventional wisdom may suggest that feature extractions based on the deepest layer i.e. “fc8” compared to “fc6” will result in the best recognition performance but our results have proved it otherwise for the two considered models. Our experiments have revealed that for the two models under consideration, the “fc6” based feature vectors have achieved the best recognition performance. State-of-the-Art recognition performances of 91.17% and 91.35% have been achieved by utilizing the “fc6” based feature vectors for the VGG_16 and VGG_19 models respectively. The recognition performance has been achieved by considering 30 sample images per class whereas the proposed system is capable of achieving improved performance by considering all sample images per class. Our research shows that for feature extraction based on CNNs, multiple layers should be considered and then the best layer can be selected that maximizes the recognition performance.

自适应特征融合的多模态实体对齐研究

多模态数据间交互式任务的兴起对于综合利用不同模态的知识提出了更高的要求,因此融合不同模态知识的多模态知识图谱应运而生.然而,现有多模态知识图谱存在图谱知识不完整的问题,严重阻碍对信息的有效利用.缓解此问题的有效方法是通过实体对齐进行知识图谱补全.当前多模态实体对齐方法以固定权重融合多种模态信息,在融合过程中忽略不同模态信息贡献的差异性.为解决上述问题,设计一套自适应特征融合机制,根据不同模态数据质量动态融合实体结构信息和视觉信息.此外,考虑到视觉信息质量不高、知识图谱之间的结构差异也影响实体对齐的效果,本文分别设计提升视觉信息有效利用率的视觉特征处理模块以及缓和结构差异性的三元组筛选模块.在多模态实体对齐任务上的实验结果表明,提出的多模态实体对齐方法的性能优于当前最好的方法.

基于联合深度统计特征对齐的鱼类目标识别方法

水下鱼类目标识别技术是认识海洋、经略海洋、向海图强的重要技术之一.基于深度学习的水下目标识别技术已成为研究热点,但是针对水下鱼类数据小样本甚至零样本识别性能亟待提高.本文基于迁移学习,提出了联合深度统计特征对齐(Joint Deep Statistical Feature Alignment, JDSFA)方法,解决小样本下的鱼类目标识别问题.以ResNet-50作为骨干网络,将均方和协方差纳入权重选择算法用来构建自适应损失函数,对齐源域和目标域之间的特征分布,联合源域损失与领域间的自适应损失,设计全局损失函数,建立深度学习识别模型,实现鱼类目标识别任务.利用公开的水下鱼类数据集QUT进行实验验证,相比目前代表性的DADAN、PMTrans、DSAN方法,JDSFA方法的鱼类识别性能分别提升了3.59%、4.96%、5.91%,结果表明了本文JDSFA方法的有效性,并对鱼类目标识别具有良好的应用价值.

结合坐标Transformer的轻量级人体姿态估计算法

针对现有的大多数自底向上人体姿态估计算法存在模型规模大、计算成本高及对边缘设备不友好等问题,提出了一种基于YOLOv5s6-Pose的轻量级多人姿态估计网络模型YOLOv5s6-Pose-CT。该模型在颈部网络中引入空间和通道重建卷积,以减少空间和通道维度上的特征冗余。同时,提出了一种坐标Transformer嵌入于主干网络中,使模型专注于长距离依赖和拥有高效的局部特征提取能力。其次,通过使用无偏特征位置对齐来解决多尺度融合过程中出现的特征错位问题。最后,使用损失函数MPDIoU对边界框的回归损失重新定义。在COCO 2017数据集上的实验结果表明,本文优化的网络模型与主流的轻量级网络EfficientHRNet-H1模型相比,在保持相同精度的同时,参数量和计算量分别减少16.2%和66.1%。相比于基准模型YOLOv5s6-Pose,参数量减少11.2%,计算量降低5.8%,平均检测精度和平均召回率分别提升2.5%和2.6%。

迁移学习在机械设备故障诊断领域的进展研究

迁移学习是一种新兴的机器学习方法,通过运用已学习的知识对不同但相关领域问题进行求解,能够较为有效的解决模型泛化能力弱、样本数据不足等问题。针对迁移学习在机械设备故障诊断领域的应用方法进行了综述,总结了三类关于迁移学习的诊断预测方法,并对迁移学习在故障诊断领域的未来研究方向进行了探讨。

基于ResCNN-BiGRU的四川方言语音识别

由于基于深度卷积神经网络的语音识别模型中缺乏对特定方言音素特征的提取能力,造成方言发音底层特征部分信息丢失,进而导致方言识别准确率不高、鲁棒性差等问题。针对上述问题,提出一种结合残差网络(RestNet)和双向门控循环网络(BiGRU)的模型,该模型以GFCC特征图为输入,同时在残差网络中设计多尺度卷积模块,通过不同大小的卷积核提取特征,然后使用双向门控循环网络捕捉序列数据中的长期依赖关系,最后采用连接时序分类算法进行标签软对齐,实现四川方言语音识别模型。在四川方言语料库上的实验结果表明,提出的模型识别性能优于现有基准模型。

深度语义关联学习的基于图像视觉数据跨域检索

基于图像的视觉数据跨域检索任务旨在搜索与输入图像在语义上一致或外形上相似的跨域图像和三维模型数据,其面临的主要问题是处理跨域数据之间的模态异质性。现有方法通过构建公共特征空间,采用域适应算法或深度度量学习算法实现跨域特征的域对齐或语义对齐,其有效性仅在单一类型的跨域检索任务中进行了验证。提出一种基于深度语义关联学习的方法,以适用多种类型的基于图像的跨域视觉数据检索任务。首先,使用异构网络提取跨域数据的初始视觉特征;然后,通过构建公共特征空间实现初始特征映射,以便进行后续的域对齐和语义对齐;最后,通过域内鉴别性学习、域间一致性学习和跨域相关性学习,消除跨域数据特征之间的异质性,探索跨域数据特征之间的语义相关性,并为检索任务生成鲁棒且统一的特征表示。实验结果表明,该方法在TU-Berlin、IM2MN和MI3DOR数据集中的平均精度均值(mAP)分别达到0.448、0.689和0.874,明显优于对比方法。

基于语义一致性约束与局部-全局感知的多模态3D视觉定位

3D多模态数据稀缺,使得传统方法进行监督训练时文本与视觉特征缺乏语义一致性。同时传统方法还易忽视局部关系与全局信息,从而导致性能不佳。针对上述问题,提出了一种基于语义一致性约束与局部-全局感知的多模态3D视觉定位方法。首先,该方法通过蒸馏2D预训练视觉语言模型知识,帮助3D模型提取到点云-文本语义一致性特征;其次设计了局部-全局感知模块,不断补充增强候选目标特征,以更精确匹配目标。在现有的3D视觉定位数据集ScanRefer上进行的实验表明,该方法在Acc@0.25 IoU和Acc@0.5 IoU两个指标上分别达到了50.53%和37.67%,超越了现有大多数3D视觉定位算法,证实了该方法的有效性。

融合注意力机制的DeeplabV3+服装图像分割方法

针对在服装图像语义分割中存在由服装颜色、纹理、背景以及多目标遮挡导致的边缘分割粗糙和分割精度低等问题,文中基于Deeplabv3+框架,提出了一种图像语义分割算法(FFDNet)。首先,模型的骨干网络采用ResNet101网络,并添加通道空间注意力模块(Feature-Enhanced Attention Module,FEAM),通过对特征图加权来挖掘并增强特征信息,提高网络表达能力。其次引入特征对齐模块(Feature Align Module,FAM)作为一种新的上采样方式,解决不同尺度特征融合之间特征未对齐导致分割错误且效率低的问题,以此提高对服装图像分割的准确性和鲁棒性。最后,FFDNet在Deepfashion2和PASCAL VOC 2012数据集上的平均交并比分别达到55.2%和79.4%;在参数量方面,该模型相比原模型在Deepfashion2上仅增加了0.61 MB。与其他现有经典模型对比,其分割性能更优,能有效捕获图像局部细节信息,减少像素分类错误。

一种基于多层特征对齐的知识蒸馏方法

实时目标检测算法(如YOLO)是为在资源有限的边缘设备上高效执行物体检测任务而设计的。因检测性能有限,提出一种基于多层特征对齐的知识蒸馏方法。为有效保留原始数据中的知识,引入将教师和学生模型的多个中间层知识纳入其中的蒸馏指标,根据训练过程中教师模型和学生模型中间层特征的差异,纳入了对齐加权因子。该方法能让学生模型从教师模型的中间层学到更多有用的知识。利用提炼出的知识对现有模型进行了增量训练,避免训练多个独立模型的资源开销。通过不同场景和条件下的实验比较,该方法在降低模型计算和存储成本的同时,有效提高目标识别的准确性。实验分析表明,在YOLO模型基础上提出的多层特征对齐蒸馏算法经COCO2017数据集验证,将学生模型的检测精度从33.3提升到40.7,有效提高模型的检测精度。

语义流引导采样结合注意力机制的脑肿瘤图像分割

U型网络结构的脑肿瘤自动分割方法由于多次卷积和采样操作会造成信息损失,导致分割效果不佳。为解决这一问题,提出了能够利用语义信息流引导上采样特征恢复的特征对齐单元,并在此基础上设计轻量级的双重注意力特征对齐网络(DAFANet)。首先,将特征对齐单元分别引入3D UNet、DMFNet和HDCNet三个经典网络,以验证其有效性和泛化性。其次,在DMFNet基础上构造轻量级的双重注意力特征对齐网络DAFANet,利用特征对齐单元强化上采样过程中的特征恢复,3D期望最大化注意力机制同时作用于特征对齐路径和级联路径,用于重点获取上下文的全程依赖关系。同时使用广义Dice损失函数提升数据不平衡时的分割精度并加快模型收敛。最后,在BraTS2018和BraTS2019公开数据集进行验证,文中所提算法在ET,WT和TC区域的分割精度分别达到80.44%,90.07%,84.57%和78.11%,90.10%,82.21%。相较于当前流行的分割网络,具有对增强肿瘤区域更好的分割效果,更擅长处理细节和边缘信息。

基于特征点匹配的排水管道声点云模型配准算法

三维重建技术逐渐成为获取全面、完备、准确的排水管道信息的关键手段;而实际检测受到管道堵塞等工况与管道检测规程等因素限制,造成所获得的管道声呐点云模型会出现位姿不同、部分重叠或空缺等情况,需要通过配准获取完整管道模型;同时,传统ICP算法针对管道模型存在效率低、精度差的问题;因此,文章提出基于特征点匹配的粗配准与改进的ICP精细配准相结合的点云配准算法;首先,利用ISS特征点检测法检测出模型特征点,通过FPFH对特征点进行进一步的描述;其次,采用RANSAC算法筛选出正确特征匹配点集,利用四元数法解算出初始变换参数完成粗配准;最后,在粗配准基础上,通过改进最近对应点查询的ICP算法完成精细配准;实验结果表明了该文算法的可行性与优越性,能为后续排水管道缺陷检测提供高完备、全面、准确的点云模型。

基于改进UperNet的结球甘蓝叶球识别方法

[目的/意义]叶球是结球甘蓝的重要部分,其生长发育对田间管理至关重要。针对叶球分割识别存在大田背景复杂、光照不均匀和叶片纹理相似等问题,提出一种语义分割算法UperNet-ESA,旨在能快速、准确地分割田间场景中结球甘蓝的外叶和叶球,以实现田间结球甘蓝的智能化管理。[方法]首先,采用统一感知解析网络(Unified Perceptual Parsing Network,UperNet)作为高效语义分割框架,将主干网络改为先进的ConvNeXt,使得模型在提升分割精度的同时也能具有较低的模型复杂度;其次,利用高效通道注意力机制(Efficient Channel Attention,ECA)融入特征提取网络的各阶段,进一步捕捉图像的细节信息;最后,通过将特征选择模块(Feature Selection Model,FSM)和特征对齐模块(Feature Alignment Model,FAM)集成到特征金字塔框架中,得到更为精确的目标边界预测结果。[结果和讨论]在自制结球甘蓝图像数据集上进行实验,与目前主流的UNet、PSPNet和DeeplabV3+语义分割模型相比,改进UperNet方法的平均交并比为92.45%,平均像素准确率为94.32%,推理速度为16.6 f/s,能够达到最佳精度-速度平衡效果。[结论]研究成果可为结球甘蓝生长智能化监测提供理论参考,对甘蓝产业发展具有重要的应用前景。

自适应特征融合的迭代实体对齐方法

针对知识图谱实体对齐任务中缺乏训练数据以及长尾实体对齐准确率较低的问题,提出一种基于自适应特征融合策略的迭代实体对齐方法,并设计一种迭代策略自动扩充训练数据的规模.该方法使用知识图谱的结构信息,并利用关系、属性和实体名称信息作为语义信息辅助对齐,从而提升对齐效果.在数据集上的实验结果表明,该模型在知识图谱实体对齐任务中效果良好.

单源域泛化中一种基于域增强和特征对齐的元学习方案

基于元学习的单源域泛化(single domain generalization,SDG)已成为解决领域偏移问题的有效技术之一。然而,源域和增强域的语义信息不一致以及域不变特征和域相关特征难以分离,使SDG模型难以实现良好的泛化性能。针对上述问题,提出了一种单源域泛化中基于域增强和特征对齐的元学习方案(meta-learning based on domain enhancement and feature alignment,MetaDefa)。利用背景替换和视觉损害技术为每一张图像生成多样且有效的增强图像,保证了源域和增强域之间的语义信息一致性;多通道特征对齐模块通过关注源域和增强域特征空间之间的相似目标区域和抑制非目标区域的特征表示充分挖掘图像信息,进而有效地提取充足的可迁移性知识。通过实验评估,MetaDefa在office-Caltech-10、office31和PACS数据集上分别取得了88.87%、73.06%和57.06%的精确度。结果表明,MetaDefa方法成功实现了源图像和增强图像之间的语义一致性和对域不变特征的充分提取,从而显著提升了单源域泛化模型的泛化性能。