Showing papers on "Feature (machine learning) published in 2022"
TL;DR: A novel deep neural network based on bidirectional-convolutional long short-term memory (BiConvLSTM) networks to determine the type, location, and direction of planetary gearbox faults by extracting spatial and temporal features from both vibration and rotational speed measurements automatically and simultaneously.
72 citations
TL;DR: An open set fault diagnosis method is proposed to address the fault diagnosis problem in a more practical scenario where the test label set consists of a portion of the training label set and some unknown classes.
Abstract: Existing data-driven fault diagnosis methods assume that the label sets of the training data and test data are consistent, which is usually not applicable for real applications since the fault modes that occur in the test phase are unpredictable. To address this problem, open set fault diagnosis (OSFD), where the test label set consists of a portion of the training label set and some unknown classes, is studied in this article. Considering the changeable operating conditions of machinery, OSFD tasks are further divided into shared-domain open set fault diagnosis (SOSFD) and cross-domain open set fault diagnosis (COSFD) in this article. For SOSFD, 1-D convolutional neural networks are trained for learning discriminative features and recognizing fault modes. For COSFD, due to the distribution discrepancy between the source and target domains, the deep model needs to learn domain-invariant features of shared classes and separate features of outlier classes. Thus, by utilizing the output of an additional domain classifier, a model named bilateral weighted adversarial networks is proposed to assign large weights to shared classes and small weights to outlier classes during the feature alignment. In the test phase, samples are classified according to the outputs of the deep model and unknown-class samples are rejected by the extreme value theory model. Experimental results on two bearing datasets demonstrate the effectiveness and superiority of the proposed method.
67 citations
TL;DR: In-situ data collected from a Singapore project (stacked twin bored tunnels) was used to prove the superiority of the proposed constrained dense convolutional autoencoder and DNN-based semi-supervised method.
40 citations
TL;DR: In this paper, a review of the recent research landscape in biometric finger vein recognition systems is presented, focusing on manuscripts related to keywords "Finger Vein Authentication System", "Anti-spoofing or Presentation Attack Detection", "Multimodal Biometric Finger Vein authentication", and their variations in four main digital research libraries such as IEEE Xplore, Springer, ACM, and Science Direct.
33 citations
TL;DR: In this paper, a self-supervised pretext task was proposed to learn a powerful supervisory signal for unsupervised representation learning, and a new teacher-student semisupervised consistency paradigm was introduced to learn to effectively transfer the pretrained representations to downstream tasks based on prediction consistency with the task-specific unlabeled data.
30 citations
TL;DR: A novel unsupervised MDA-based transfer learning approach called multisource domain factorization network (MDFN) is proposed in this paper, where the generalized diagnosis knowledge is learned from multiple sources and then used for diagnosing the target task.
26 citations
TL;DR: In this article, a method based on Raman spectroscopy combined with generative adversarial network and multiclass support vector machine was proposed to classify foodborne pathogenic bacteria. But it is still a challenge to overcome the cumbersome culture process of bacteria and the need for a large number of samples, which hinder qualitative analysis, to obtain a high classification accuracy.
20 citations
TL;DR: The few-shot learning is introduced to construct a deep learning model named Convolutional Relation Network (CRN) for FER in the wild, which is learnt by exploiting a feature similarity comparison among the sufficient samples of the emotion categories to identify new classes with few samples.
Abstract: Recent deep learning based facial expression recognition (FER) methods are mostly driven by the availability of large amount of training data. However, availability of such data is not always possible for FER in the wild where the infeasibility of obtaining sufficient training samples for each emotion category. Therefore, in this paper, we introduce the few-shot learning to construct a deep learning model named Convolutional Relation Network (CRN) for FER in the wild, which is learnt by exploiting a feature similarity comparison among the sufficient samples of the emotion categories to identify new classes with few samples. Specifically, our method learns a metric space in which classification can be performed by computing distances to capitalize on powerful discriminative ability of deep expression features to generalize the predictive power of the network. To achieve this, the features are constrained to maximize the distance between the features of different classes and discover the commonality of the same classes. Extensive experiments on three challenging in-the-wild datasets demonstrate that the proposed model significantly outperforms state-of-the-art methods.
19 citations
TL;DR: The Unsupervised Descriptor Selection (UDS) is proposed to tackle few-shot learning tasks, and achieves the comparable performance to state-of-the-art methods, and improves the performance of prior meta-learning methods.
16 citations
TL;DR: In this paper, a robust approach based on Sobol sensitivity analysis is proposed to improve the robustness of support vector machine (SVM) models to the impact of feature uncertainties.
Abstract: This paper addresses the problem of classification when target data are subject to feature uncertainties. A robust approach based on Sobol sensitivity analysis is proposed to improve the robustness of support vector machine (SVM) models. SVM is a supervised machine learning method for pattern recognition whose performance depends on the definition of its hyperparameters and the quality of data. The proposed approach analyzes the impact of the uncertainties on the predictive performance of SVM based on Sobol’ sensitivity analysis. Afterwards, a new parameter is introduced to improve the robustness of SVM to the impact of uncertainties. The efficiency of this approach is evaluated by applying it to six real-world datasets. The results are then discussed and analyzed.
16 citations
01 Jan 2022
TL;DR: In this article, the authors proposed a hybrid method for face recognition using Principal Component Analysis (PCA), Discrete Wavelet Transform (DWT), Independent Component Analysis and Support Vector Machine (SVM).
Abstract: Face recognition need is fine assured as enormous industrial relevance use them to implement one or another objective. As the programmes move closer to everyday usage to hold a database of actual events, an individual’s identification primarily demanded as an instance of consistency. As facial recognition has beating advantages over other industrial applications and human eyes can quickly evaluate performance, improved algorithms and smaller computing costs are continuously improving this methodology. This research takes the conventional algorithms of recognition in the first stage and uses hybrid approaches to counter their limitations. The study starts with basic computation of global face features using Principal Component Analysis (PCA), Discrete Wavelet Transform (DWT) and Independent Component Analysis (ICA), with some standard classifiers like Neural Network (NN) and Support Vector Machine (SVM). As the learning rate is high in machine learning, then the system’s accuracy goes high, but increases the area and cost overhead. Fusion-based methods have been proposed in further work to overcome that training limitation, based on Harris corner, Speed up Robust Features (SURF) and DWT + PCA system model where only 10% training sample has been taken on Essex database, and 99.45% accuracy is achieved. Creating the Fusion rule requires some hit and trial methods that may not be Universal in every database. To overcome this limitation further an efficient Hybrid method proposed which elaborates the local features Linear Binary Pattern (LBP), Histogram Oriented Gradients (HOG), Gabor wavelet and global features (DWT, PCA) of the face. Further, this feature trained with Neural Network classifier to obtained better accuracy nearly 99.40% with single image training from each class.
TL;DR: Wang et al. as mentioned in this paper proposed Feature Refinement (FeatRef) with expression-specific feature learning and fusion for micro-expression recognition that aims to obtain salient and discriminative features for specific expressions and predicts expressions by fusing expression specific features.
TL;DR: To search for the indispensable and representative features of each label-related subset, the local-reduction-based feature selection method (LRFS-α) is designed and comprehensive experimental results on fifteen multi- label datasets characterize the performance of the methods against other seven multi-label learning methods.
TL;DR: An algorithm for regression analysis that addresses features typical for big data sets, which is called “sparse shooting S”, meaning that many of the regression coefficients are set to zero, hereby selecting the most relevant predictors.
TL;DR: A novel two-step dense attention mechanism to discover attribute-guided local visual features for both EL and FS categories in an integrated manner for fine-grained GZSL and demonstrates that the proposed method outperforms contemporary methods on benchmark datasets.
TL;DR: In this article, the authors demonstrate that an appropriate pre-training of a CNN model in the task of handwritten text-based writer identification task, can dramatically improve the efficiency of the CNN in the OSV task, enabling to obtain state-of-the-art performance with an order of magnitude less training signature samples.
Abstract: Handwritten signature is a common biometric trait, widely used for confirming the presence or the consent of a person. Offline Signature Verification (OSV) is the task of verifying the signer using static signature images captured after the finish of signing process, with many applications especially in the domain of forensics. Deep Convolutional Neural Networks (CNNs) can generate efficient feature representations, but their training is data-intensive. Since limited training data is an intrinsic problem of an OSV system’s development, this work focuses on addressing the problem of learning informative features by employing prior knowledge from a similar task in a domain with an abundance of training data. In particular, we demonstrate that an appropriate pre-training of a CNN model in the task of handwritten text-based writer identification task, can dramatically improve the efficiency of the CNN in the OSV task, enabling to obtain state-of-the-art performance with an order of magnitude less training signature samples. In the proposed scheme, after the pre-training of the CNN in writer identification task through specially processed handwritten text data, the learned features are tailored to the signature problem though a metric learning stage that utilizes contrastive loss to learn a mapping of the signatures’ features to a latent space that suits the OSV task. At the final stage, the proposed scheme utilizes Writer-Dependent (WD) classifiers learned on a few reference samples from each writer. Our system is tested on the three challenging signature datasets, CEDAR, MCYT-75 and GPDS300GRAY. The obtained accuracy in terms of Equal Error Rates (EER) is statistically equivalent to the popular SigNet CNN, despite a significantly smaller training set of signature images and no use of skilled forgeries signatures during training.
TL;DR: A novel lightweight 3-D place recognition and fast sequence matching method, capable of recognizing places from a previous trajectory regardless of viewpoints and temporary observation differences, is proposed, which outperforms the relative state of the art.
Abstract: Recognizing the same place undervariant viewpoint differences is the fundamental capability for human beings and animals However, such a strong place recognition ability in robotics is still an unsolved problem Extracting local invariant descriptors from the same place under various viewpoint differences is difficult This article seeks to provide robots with a human-like place recognition ability using a new 3-D feature learning method This article proposes a novel lightweight 3-D place recognition and fast sequence matching to achieve robust 3-D place recognition, capable of recognizing places from a previous trajectory regardless of viewpoints and temporary observation differences Specifically, we extracted the viewpoint-invariant place feature from 2-D spherical perspectives by leveraging spherical harmonics’ orientation-equivalent property To improve sequence-matching efficiency, we designed a coarse-to-fine fast sequence-matching mechanism to balance the matching efficiency and accuracy Despite the apparent simplicity, our proposed approach outperforms the relative state of the art In both public and self-gathered datasets with orientation/translation differences or noise observations, our method can achieve above 95% average recall for the best match with only 18% inference time of PointNet-based place recognition methods
TL;DR: Wang et al. as mentioned in this paper proposed Discriminative Adversarial Domain Generalization (DADG) with meta-learning-based cross-domain validation to learn a domain-invariant feature representation from source domains and generalize it to the unseen domains.
TL;DR: In this article, Tensor Factorized Neural Networks (TFNN) and Attention Gated Tensor factorized Neural Network (AG-TFNN), were used for speech emotion recognition. But, their performance was not as good as CNN and LSTM.
TL;DR: An Autoencoder(AE)-based feature construction approach to remove the dependency of manually correlating commands and generate an efficient representation by automatically learning the semantic similarity between input features extracted through commands data, which resulted in providing meaningful clustering interpretations.
TL;DR: In this article, a multimodal emotion classification framework that uses multichannel physiological signals and introduces two key techniques, hybrid feature extraction and adaptive decision fusion, was proposed for emotion classification.
TL;DR: A Fault Detection and Diagnosis framework for Non-Linear Processes utilizing Dynamic Neural Networks and feature reduction methods is proposed and it is demonstrated that this method outperforms state of the art methods in the majority of those faults.
TL;DR: In this paper, a novel Discrete Wavelet Concatenated Mesh Tree (DW-CMT) and ternary chess pattern (TCP) based ECG signal recognition method is presented.
TL;DR: Wang et al. as mentioned in this paper proposed a robust multi-feature collective non-negative matrix factorization (RMCNMF) model to handle noise and sample variation in ECG Biometrics.
TL;DR: In this paper, a novel Weight-based Meta Metric Learning (W2ML) method is proposed for accurate open-set touchless palmprint recognition, where only a part of categories is seen during training.
TL;DR: Zhang et al. as discussed by the authors proposed an identity-guided human region segmentation (HRS) method for person retrieval, which learns a set of distinct region bases that are consistent across a given dataset, and predicts informative region segments by grouping intermediate feature vectors based on their similarity to these bases.
TL;DR: In this article, a subspace alignment extreme learning machine (SAELM) is proposed to learn a robust subspace to increase the consistency between domains and enhance the feature-label dependency of the source domain.
Abstract: The drift caused by gas sensors has always been a bottleneck in the development of electronic nose (E-nose) systems. Traditional drift compensation methods directly correct the drift components, making such approaches time-consuming and laborious. In the field of E-nose drift compensation, cross-domain adaption learning is an efficient technique. In this paper, we propose a novel subspace alignment extreme learning machine (SAELM) that considers multiple criteria to construct a unified extreme learning machine (ELM)-based feature representation space and thus achieve domain alignment. First, the method minimizes both the geometric and statistical distributions between different domains. Second, the dependence between features and labels is enhanced using the Hilbert–Schmidt independence criterion (HSIC) to alleviate the blurring of the correspondence between the two caused by drift. Third, to improve the feature extraction ability of the subspace learning method, the l 2,1 norm is leveraged to constrain the output weights of the ELM. The aim of this method is to learn a robust subspace to increase the consistency between domains and enhance the feature–label dependency of the source domain while preserving the intrinsic information of both domains. Extensive experiments on sensor drift data are conducted, and the proposed SAELM method yields the greatest improvements on E-nose drift datasets.
01 Jan 2022
TL;DR: In this article, the proposed neural network (NN) model with the different features like MFCC and mel spectrogram extracted from the speech signal to recognize the gender from voice is considered as one of the essential tasks to be detected for such applications.
Abstract: The human speech contains paralinguistic information used in many speech recognition applications like automatic speech recognition, speaker recognition, and verification. Gender from voice is considered as one of the essential tasks to be detected for such applications. To build a model from a training set, a set of relevant speech features is extracted in order to distinguish gender (i.e., female or male) from a speech signal. This paper focuses on comparison of the proposed neural network (NN) model with the different features like MFCC and mel spectrogram extracted from the speech signal to recognize the gender. Experiments are carried on Mozilla voice dataset and evaluated performance of the network. Experiments show that the combination of MFCC and mel feature sets shows the better accuracy with 94.32%.
TL;DR: Wang et al. as discussed by the authors investigated staking-based ensemble learning on safety-related properties to assist the risk assessment based on molecular structure-based features, individual and ensemble models are built and compared using heterogeneous machine learning methods.
TL;DR: Wang et al. as discussed by the authors proposed a classification and recognition method based on signal enhancement, and the experimental results on the public database show that the proposed method can effectively realize classification and classification in the environment of small sample EEG signals.