基于深度置信网络与信息融合的齿轮故障诊断方法

针对齿轮在复杂运行工况下故障特征提取困难,传统故障诊断方法的识别精度易受人工提取特征的影响,以及单传感器获取信息不全面等问题,提出基于深度置信网络(DBN)与信息融合的齿轮故障诊断方法。通过多传感器信息融合技术对每个传感器采集的振动信号进行数据层融合;利用DBN进行自适应特征提取从而实现故障分类。为了避免因人为选择DBN结构参数,导致模型识别精度下降的问题,利用改进的混合蛙跳算法(ISFLA)对DBN结构参数进行优化。试验表明,与BP神经网络、未经优化的DBN以及单传感器故障诊断相比,该研究提出的信息融合及优化方法具有更高的故障识别精度。     

基于卷积神经网络的印刷套准识别方法

目的 针对目前印刷套准识别方法依赖于经验人工设计特征提取的问题,提出一种不需要人工提取图像特征的卷积神经网络模型,实现印刷套准状态的识别.方法 采用图像增强技术实现不均衡训练集的均衡化,增加训练集图像的数量,提高模型的识别准确率.设计基于AlexNet网络结构的印刷套准识别模型的结构参数,分析批处理样本数量和基础学习率对模型性能的影响规律.结果 文中方法获得的总印刷套准识别准确率为0.9860,召回率为1.0000,分类准确率几何平均数为0.9869.结论 文中方法能自动提取图像特征,不依赖于人工设计的特征提取方法.在构造的数据集上,文中方法的分类性能优于实验中的支持向量机方法.     

A Novel Combinational Convolutional Neural Network for Automatic Food-Ingredient Classification

With the development of deep learning and Convolutional Neural Networks (CNNs), the accuracy of automatic food recognition based on visual data have significantly improved. Some research studies have shown that the deeper the model is, the higher the accuracy is. However, very deep neural networks would be affected by the overfitting problem and also consume huge computing resources. In this paper, a new classification scheme is proposed for automatic food-ingredient recognition based on deep learning. We construct an up-to-date combinational convolutional neural network (CBNet) with a subnet merging technique. Firstly, two different neural networks are utilized for learning interested features. Then, a well-designed feature fusion component aggregates the features from subnetworks, further extracting richer and more precise features for image classification. In order to learn more complementary features, the corresponding fusion strategies are also proposed, including auxiliary classifiers and hyperparameters setting. Finally, CBNet based on the well-known VGGNet, ResNet and DenseNet is evaluated on a dataset including 41 major categories of food ingredients and 100 images for each category. Theoretical analysis and experimental results demonstrate that CBNet achieves promising accuracy for multi-class classification and improves the performance of convolutional neural networks.     

基于人工鱼群算法-极限学习机的多手势精准识别

手势作为人机交互的重要方式,因灵活性与便捷性强,已成为控制领域的研究重点。针对上肢康复机器人手势识别技术存在的不足,结合特征组合与滑动窗口法,提出一种基于人工鱼群算法（artificial fish swarm algorithm,AFSA）优化的极限学习机（extreme learning machine,ELM）的多手势精准识别方法,以提高手势识别的准确率。首先,运用表面肌电测量系统采集人体常用的8种手势对应的表面肌电信号（surface electromyography,SEMG）,作为后续分类模型的信号源,并运用去噪技术与起止点检测技术对SEMG进行预处理;然后,选取通过主成分分析（principal components analysis,PCA）降维处理后的最优特征组合与最优滑动窗口;接着,采用AFSA搜寻ELM的最优输入权值和隐含阈值,以提高其分类准确率;最后,对AFSA优化的ELM（AFSA-ELM）分类模型、反向传播（back propagation,BP）神经网络分类模型和未优化的ELM分类模型进行比较,以验证所提出方法的精准性。实验结果表明,结合最优特征组合与最优滑动窗口设计的AFSA-ELM分类模型对多种手势的平均识别准确率高达97.4%,比BP神经网络分类模型和未优化的ELM分类模型分别高3.5%和1.6%,验证了所提出方法的识别精准性。研究结果可为手势识别提供新思路,进而为人体上肢动作的深度分析和上肢康复机器人手势识别算法的优化提供理论基础和参考。     

基于特征融合和B-SVM的鸟鸣声识别算法

为了实现在野外通过低成本嵌入式系统识别鸟类,提出了基于特征融合和B-SVM的鸟鸣声识别方法。对鸟鸣声信号提取梅尔频率倒谱系数、翻转梅尔频率倒谱系数、短时能量和短时过零率组成特征参数,通过线性判别算法对特征参数进行特征融合。利用黑寡妇算法通过测试集对支持向量机模型的核参数和损失值进行优化得到B-SVM模型。利用Xeno-canto鸟鸣声数据集对本文算法进行了测试,结果表明该方法的识别准确率为93.23%。算法维度参数的大小和融合特征维度的高低是影响算法识别效果的重要因素。在相同条件下,文中所提的基于特征融合和B-SVM模型的鸟鸣声识别算法相较于其他特征参数和模型,识别的准确率更高,为野外鸟类识别提供了参考。     

基于线性判别分析的决策融合脑电意识动态分类

脑电信号的识别与分类是脑机接口技术的热点研究问题,单一分类器不能很好利用特征以及分类器的适应性,导致识别的准确率很难进一步提高,基于线性判别分析的分类决策级融合策略,可用于提高脑-机接口系统的分类准确率。首先,通过分离出两种分类器的假性试验特征,从这两种方法中选择更有可能正确决策提高分类准确性;其次为了测量每个决策的不确定性,使用与所对应分类器的最大和第二大相关系数提取特征向量。基于这一思想,提出了一种新的决策选择器,该方法通过整合两种基于线性判别分析的算法选择更有可能是准确的决策,从而达到提高脑电信号分类准确度。实验结果表明,该方法通过与精度相近的算法相结合在运动想象数据分类上获得了较好的分类准确率。     

基于多分辨率时频特征融合的声学场景分类

声学场景分类是计算机听觉中最难的任务之一,在单一特征条件下采用基本的卷积神经网络相对于传统的分类方法精度已经有所提升,但是效果依然不够理想。针对这一问题,在卷积神经网络框架下,提出了一种基于时频特征融合的声学场景分类方案。在分类模型构建方面,提出一种多分辨率卷积池化方案,构造多分辨率卷积神经网络,以更好地适应提取特征的时频结构;在特征选取方面,融合低层次包络特征对数——Mel子带能量和高层次结构特征——非负矩阵分解系数矩阵,把两种二维特征堆叠为三维特征送入分类模型。在2017年和2018年声学场景分类和事件检测挑战赛的开发数据集上进行了训练和测试。实验结果表明,文中提出方案比基线系统的分类精度分别提高7.5%和10.3%,可有效改善分类效果。     

基于深度学习的舰船辐射噪声多特征融合分类

为了改善分类系统的性能,进一步提高舰船辐射噪声分类的正确率,该文提出了一种基于深度神经网络的多特征融合分类方法。该方法首先提取舰船辐射噪声几种不同的特征,将提取的特征同时用于训练具有多个输入分支的深度神经网络,使网络直接在多种特征参数上进行联合学习,通过神经网络的输入分支和连接层实现特征融合,再对舰船辐射噪声进行分类。为了特征深度学习提取了舰船辐射噪声的频谱特征、梅尔倒谱系数和功率谱特征,并将多特征融合分类方法与在一种特征上进行深度学习分类方法的正确率进行对比。实验结果表明,基于深度学习的多特征融合分类方法可以有效地提高舰船辐射噪声分类的正确率,是一种可行的分类方法。     

基于3D和1D多特征融合的语音情感识别算法

针对语音情感识别任务中特征提取单一、分类准确率低等问题,提出一种3D和1D多特征融合的情感识别方法,对特征提取算法进行改进.在3D网络,综合考虑空间特征学习和时间依赖性构造,利用双线性卷积神经网络(Bilinear Convolutional Neural Network,BCNN)提取空间特征,长短期记忆网络(Sho...     

Research on Facial Expression Capture Based on Two-Stage Neural Network

To generate realistic three-dimensional animation of virtual character, capturing real facial expression is the primary task. Due to diverse facial expressions and complex background, facial landmarks recognized by existing strategies have the problem of deviations and low accuracy. Therefore, a method for facial expression capture based on two-stage neural network is proposed in this paper which takes advantage of improved multi-task cascaded convolutional networks (MTCNN) and high-resolution network. Firstly, the convolution operation of traditional MTCNN is improved. The face information in the input image is quickly filtered by feature fusion in the first stage and Octave Convolution instead of the original ones is introduced into in the second stage to enhance the feature extraction ability of the network, which further rejects a large number of false candidates. The model outputs more accurate facial candidate windows for better landmarks recognition and locates the faces. Then the images cropped after face detection are input into high-resolution network. Multi-scale feature fusion is realized by parallel connection of multi-resolution streams, and rich high-resolution heatmaps of facial landmarks are obtained. Finally, the changes of facial landmarks recognized are tracked in real-time. The expression parameters are extracted and transmitted to Unity3D engine to drive the virtual character's face, which can realize facial expression synchronous animation. Extensive experimental results obtained on the WFLW database demonstrate the superiority of the proposed method in terms of accuracy and robustness, especially for diverse expressions and complex background. The method can accurately capture facial expression and generate three-dimensional animation effects, making online entertainment and social interaction more immersive in shared virtual space.     

Accurate Multi-Scale Feature Fusion CNN for Time Series Classification in Smart Factory

Time series classification (TSC) has attracted various attention in the community of machine learning and data mining and has many successful applications such as fault detection and product identification in the process of building a smart factory. However, it is still challenging for the efficiency and accuracy of classification due to complexity, multi-dimension of time series. This paper presents a new approach for time series classification based on convolutional neural networks (CNN). The proposed method contains three parts: short-time gap feature extraction, multi-scale local feature learning, and global feature learning. In the process of short-time gap feature extraction, large kernel filters are employed to extract the features within the short-time gap from the raw time series. Then, a multi-scale feature extraction technique is applied in the process of multi-scale local feature learning to obtain detailed representations. The global convolution operation with giant stride is to obtain a robust and global feature representation. The comprehension features used for classifying are a fusion of short time gap feature representations, local multi-scale feature representations, and global feature representations. To test the efficiency of the proposed method named multi-scale feature fusion convolutional neural networks (MSFFCNN), we designed, trained MSFFCNN on some public sensors, device, and simulated control time series data sets. The comparative studies indicate our proposed MSFFCNN outperforms other alternatives, and we also provided a detailed analysis of the proposed MSFFCNN.     

基于声发射信号EMD-WPD特征融合的航天器在轨泄漏辨识方法

长期运行在空间环境中的航天器可能由于撞击、振动、老化等因素而发生气体泄漏,在轨泄漏辨识对航天器安全保障具有重要意义.提出了一种基于声发射信号经验模态分解(empirical mode decomposition,EMD)和小波包分解(wavelet packet decomposition,WPD)特征融合的航天器泄漏...     

一种视听融合的水下目标识别方法研究

特征提取是水下目标识别研究中最为关键的技术之一,特征参数的优劣将直接决定分类识别系统的性能。将声信号的听觉与视觉感知特征结合,应用于水下目标识别,通过实验得出如下结论,相比于单独应用听觉特征,融合特征的平均识别率能提高4%~6%以上,特别是将听觉特征与声谱图的Gabor小波变换特征、灰度-梯度共生特征进行融合后,分类性能较好,平均达到87%以上。     

基于特征选择和GWO-KELM的鸟声识别算法

针对鸟声识别算法中提取特征单一、分类准确率低等问题,提出一种基于混合特征选择和灰狼算法优化核极限学习机的鸟声识别方法。首先从鸟声数据中提取大规模声学特征集ComParE,其次计算每个特征的F_score并进行排序,然后以广义顺序向前浮动搜索(Generalized Sequential Forward Floating Search, GSFFS)为搜索策略,特征子集在核极限学习机(Kernel Limit Learning Machine, KELM)上十折交叉验证的正确率,作为特征选择标准进行特征选择,得到适用于鸟声识别的特征子集,最后通过灰狼算法(Grey Wolf Optimizer, GWO)选择最优KELM参数识别鸟声。在柏林自然科学博物馆鸟声数据库中进行实验,该方法在60类鸟声识别平均正确率和F1-score达到94.45%和92.29%。结果表明,该方法相较于传统自行设计提取的单一特征集具有更高的识别精度,GWO-KELM模型比网格搜索方式更易找到全局最优值。     

一种车辆识别代号检测和识别的弱监督学习方法

车辆识别代号对于车辆年检具有重要的意义.由于缺乏字符级标注,无法对车辆识别代号进行单字符风格校验.针对该问题,设计了一种单字符检测和识别框架,并对此框架提出了一种无须字符级标注的弱监督学习方法.首先,对VGG16-BN各个层次的特征信息进行融合,获得具有单字符位置信息与语义信息的融合特征图;其次,设计了一个字符检测分支...     

Feature selection and fault‐severity classification–based machine health assessment methodology for point machine sliding‐chair degradation

In this paper, we propose an offline and online machine health assessment (MHA) methodology composed of feature extraction and selection, segmentation‐based fault severity evaluation, and classification steps. In the offline phase, the best representative feature of degradation is selected by a new filter‐based feature selection approach. The selected feature is further segmented by utilizing the bottom‐up time series segmentation to discriminate machine health states, ie, degradation levels. Then, the health state fault severity is extracted by a proposed segment evaluation approach based on within segment rate‐of‐change (RoC) and coefficient of variation (CV) statistics. To train supervised classifiers, a priori knowledge about the availability of the labeled data set is needed. To overcome this limitation, the health state fault‐severity information is used to label (eg, healthy, minor, medium, and severe) unlabeled raw condition monitoring (CM) data. In the online phase, the fault‐severity classification is carried out by kernel‐based support vector machine (SVM) classifier. Next to SVM, the k‐nearest neighbor (KNN) is also used in comparative analysis on the fault severity classification problem. Supervised classifiers are trained in the offline phase and tested in the online phase. Unlike to traditional supervised approaches, this proposed method does not require any a priori knowledge about the availability of the labeled data set. The proposed methodology is validated on infield point machine sliding‐chair degradation data to illustrate its effectiveness and applicability. The results show that the time series segmentation‐based failure severity detection and SVM‐based classification are promising.     

Feature selection and combination criteria for improving accuracy in protein structure prediction

The classification of protein structures is essential for their function determination in bioinformatics. At present, a reasonably high rate of prediction accuracy has been achieved in classifying proteins into four classes in the SCOP database according to their primary amino acid sequences. However, for further classification into fine-grained folding categories, especially when the number of possible folding patterns as those defined in the SCOP database is large, it is still quite a challenge. In our previous work, we have proposed a two-level classification strategy called hierarchical learning architecture (HLA) using neural networks and two indirect coding features to differentiate proteins according to their classes and folding patterns, which achieved an accuracy rate of 65.5%. In this paper, we use a combinatorial fusion technique to facilitate feature selection and combination for improving predictive accuracy in protein structure classification. When applying various criteria in combinatorial fusion to the protein fold prediction approach using neural networks with HLA and the radial basis function network (RBFN), the resulting classification has an overall prediction accuracy rate of 87% for four classes and 69.6% for 27 folding categories. These rates are significantly higher than the accuracy rate of 56.5% previously obtained by Ding and Dubchak. Our results demonstrate that data fusion is a viable method for feature selection and combination in the prediction and classification of protein structure.     

Real and Altered Fingerprint Classification Based on Various Features andClassifiers

Biometric recognition refers to the identification of individuals through their unique behavioral features (e.g., fingerprint, face, and iris). We need distinguishing characteristics to identify people, such as fingerprints, which are world-renowned as the most reliable method to identify people. The recognition of fingerprints has become a standard procedure in forensics, and different techniques are available for this purpose. Most current techniques lack interest in image enhancement and rely on high-dimensional features to generate classification models. Therefore, we proposed an effective fingerprint classification method for classifying the fingerprint image as authentic or altered since criminals and hackers routinely change their fingerprints to generate fake ones. In order to improve fingerprint classification accuracy, our proposed method used the most effective texture features and classifiers. Discriminant Analysis (DCA) and Gaussian Discriminant Analysis (GDA) are employed as classifiers, along with Histogram of Oriented Gradient (HOG) and Segmentation-based Feature Texture Analysis (SFTA) feature vectors as inputs. The performance of the classifiers is determined by assessing a range of feature sets, and the most accurate results are obtained. The proposed method is tested using a Sokoto Coventry Fingerprint Dataset (SOCOFing). The SOCOFing project includes 6,000 fingerprint images collected from 600 African people whose fingerprints were taken ten times. Three distinct degrees of obliteration, central rotation, and z-cut have been performed to obtain synthetically altered replicas of the genuine fingerprints. The proposal achieved massive success with a classification accuracy reaching 99%. The experimental results indicate that the proposed method for fingerprint classification is feasible and effective. The experiments also showed that the proposed SFTA-based GDA method outperformed state-of-art approaches in feature dimension and classification accuracy.