Effective Feature Extraction Method Research Articles

Feasibility study on person identification utilizing frequency response functions of human fingers

Feasibility study on person identification utilizing frequency response functions of human fingers

An efficient multiscale integrated attention method combined with hyperspectral system to identify the quality of rice with different storage periods and humidity

An efficient multiscale integrated attention method combined with hyperspectral system to identify the quality of rice with different storage periods and humidity

Head Pose Estimation Patterns as Deepfake Detectors

The capacity to create ”fake” videos has recently raised concerns about the reliability of multimedia content. Identifying between true and false information is a critical step toward resolving this problem. On this issue, several algorithms utilizing deep learning and facial landmarks have yielded intriguing results. Facial landmarks are traits that are solely tied to the subject’s head posture. Based on this observation, we study how Head Pose Estimation (HPE) patterns may be utilized to detect deepfakes in this work. The HPE patterns studied are based on FSA-Net, SynergyNet, and WSM, which are among the most performant approaches on the state of the art. Finally, using a machine learning technique based on K-Nearest Neighbor and Dynamic Time Warping, their temporal patterns are categorized as authentic or false. We also offer a set of experiments for examining the feasibility of using deep learning techniques on such patterns. The findings reveal that the ability to recognize a deepfake video utilizing an HPE pattern is dependent on the HPE methodology. On the contrary, performance is less dependent on the performance of the utilized HPE technique. Experiments are carried out on the FaceForensics++ dataset, that presents both identity swap and expression swap examples. The findings show that FSA-Net is an effective feature extraction method for determining whether a pattern belongs to a deepfake or not. The approach is also robust in comparison to deepfake videos created using various methods or for different goals. In mean the method obtain 86% of accuracy on the identity swap task and 86.5% of accuracy on the expression swap. These findings offer up various possibilities and future directions for solving the deepfake detection problem using specialized HPE approaches, which are also known to be fast and reliable.

Revealing leaf species through specific contour and region-based features extraction

Revealing leaf species through specific contour and region-based features extraction

Variational Auto-Encoder Reconstruction Networks for Classification of Hyperspectral and LiDAR Data

The classification of remote sensing scene objects has been the subject of extensive studies in recent years due to the quick advancement of earth observation and remote sensing technology. More concentration, Hyperspectral images, and LiDAR data are complementary, and their combined use for data fusion can better mine the multi-dimensional features of ground objects in remote sensing scenes, which can effectively improve the classification accuracy and reliability of ground objects in remote sensing scenes. Single-modal remote sensing data frequently cannot fully meet the needs of ground feature classification due to the increasingly complex types of ground features. In order to solve this problem, we have developed two distinct multi-source fusion classification approaches using LiDAR and hyperspectral data and deep learning techniques. They are the reconstructed multi-layer perceptron network based on the variational auto-Encoder (encoding-decoding form) and the two-stream input convolutional neural network based on the cross-channel reconstruction mechanism. These two approaches can help us find more effective and deeper feature extraction methods and feature fusion methods in this research direction and design training. We adopted the network architecture model used in this area of research and used experimental data to demonstrate the effectiveness and superiority of the suggested network model.

A novel rolling bearing fault diagnosis method based on continuous hierarchical fractional range entropy

A novel rolling bearing fault diagnosis method based on continuous hierarchical fractional range entropy

Statistically significant features improve binary and multiple Motor Imagery task predictions from EEGs.

In recent studies, in the field of Brain-Computer Interface (BCI), researchers have focused on Motor Imagery tasks. Motor Imagery-based electroencephalogram (EEG) signals provide the interaction and communication between the paralyzed patients and the outside world for moving and controlling external devices such as wheelchair and moving cursors. However, current approaches in the Motor Imagery-BCI system design require effective feature extraction methods and classification algorithms to acquire discriminative features from EEG signals due to the non-linear and non-stationary structure of EEG signals. This study investigates the effect of statistical significance-based feature selection on binary and multi-class Motor Imagery EEG signal classifications. In the feature extraction process performed 24 different time-domain features, 15 different frequency-domain features which are energy, variance, and entropy of Fourier transform within five EEG frequency subbands, 15 different time-frequency domain features which are energy, variance, and entropy of Wavelet transform based on five EEG frequency subbands, and 4 different Poincare plot-based non-linear parameters are extracted from each EEG channel. A total of 1,364 Motor Imagery EEG features are supplied from 22 channel EEG signals for each input EEG data. In the statistical significance-based feature selection process, the best one among all possible combinations of these features is tried to be determined using the independent t-test and one-way analysis of variance (ANOVA) test on binary and multi-class Motor Imagery EEG signal classifications, respectively. The whole extracted feature set and the feature set that contain statistically significant features only are classified in this study. We implemented 6 and 7 different classifiers in multi-class and binary (two-class) classification tasks, respectively. The classification process is evaluated using the five-fold cross-validation method, and each classification algorithm is tested 10 times. These repeated tests provide to check the repeatability of the results. The maximum of 61.86 and 47.36% for the two-class and four-class scenarios, respectively, are obtained with Ensemble Subspace Discriminant among all these classifiers using selected features including only statistically significant features. The results reveal that the introduced statistical significance-based feature selection approach improves the classifier performances by achieving higher classifier performances with fewer relevant components in Motor Imagery task classification. In conclusion, the main contribution of the presented study is two-fold evaluation of non-linear parameters as an alternative to the commonly used features and the prediction of multiple Motor Imagery tasks using statistically significant features.

Early fault diagnosis of rotating machinery based on composite zoom permutation entropy

• Composite zoom permutation entropy (CZPE) is proposed to extract features. • The mechanism of the composite analysis is provided to reveal the merit of CZPE. • A diagnosis framework is developed to identify early faults of rotating machinery. • The proposed method is verified by simulational and experimental results. Fault diagnosis of rotating machinery serves an important role in informing system operation and predictive maintenance decisions. To quantify the fault information from vibrational signals, the multiscale permutation entropy has become a promising tool for fault diagnosis of rotating machinery. However, multiscale permutation entropy fails to extract weak features of early faults because it can hardly capture the tiny oscillation patterns of signals over the full frequency band. To address this issue, this paper presents an effective feature extraction method called composite zoom permutation entropy. First, composite zoom permutation entropy employs multiple wavelets to capture complete fault features with multiple resolutions over the full frequency band. Then the composite analysis is performed to improve the separability of extracted features for identifying different early faults. Based on composite zoom permutation entropy, a diagnosis framework has been developed to improve the operational reliability of rotating machinery by identifying faults as early as possible. The simulation results show that composite zoom permutation entropy has better extraction ability compared with other permutation entropy based methods. The experimental results show that the proposed method outperforms existing methods in identifying early faults of rotating machinery.

An adaptive feature extraction technique for ship-radiated noise based on weighted multiscale mathematical morphological filtering

Aiming at the problem of feature extraction of ship-radiated noise in complex marine environment, an adaptive weighted multiscale mathematical morphological filtering method is proposed. The weight coefficients of the structure elements of different scales in multiscale morphological filtering are determined adaptively by an improved particle swarm optimization algorithm, and the Teager energy kurtosis is used as the evaluation index of the filtered optimal signal to provide optimized weights for each scale. Finally, the optimized multiscale mathematical morphology filter with selective adaptive weights is bound and applied to the feature extraction of ship-radiated noise. The analysis results of simulated signal and measured ship-radiated noise show that this method has strong noise suppression and feature extraction ability, and the calculation is simple and fast, which provides an effective method for ship feature extraction.

Arrhythmia Classification and Diagnosis Based on ECG Signal: A Multi-Domain Collaborative Analysis and Decision Approach

Electrocardiogram (ECG) signal plays a key role in the diagnosis of arrhythmia, which will pose a great threat to human health. As an effective feature extraction method, deep learning has shown excellent results in processing ECG signals. However, most of these methods neglect the cooperation between the multi-lead ECG series correlation and intra-series temporal patterns. In this work, a multi-domain collaborative analysis and decision approach is proposed, which makes the classification and diagnosis of arrhythmia more accurate. With this decision, we can realize the transition from the spatial domain to the spectral domain, and from the time domain to the frequency domain, and make it possible that ECG signals can be more clearly detected by convolution and sequential learning modules. Moreover, instead of the prior method, the self-attention mechanism is used to learn the relation matrix between the sequences automatically in this paper. We conduct extensive experiments on eight advanced models in the same field to demonstrate the effectiveness of our method.

Design of an intrinsically scalable model for medical disease identification using hybrid swarm intelligence with augmented transfer learning

Identification of medical diseases is a multidomain task that involves design of effective models for data collection, pre-processing, segmentation, feature extraction & reduction, with classification & post-processing. To design a medical expert system, researchers need to incorporate various input-specific processing models, that can cater to numerical, sequential, and image datasets. A wide variety of such system models are proposed by researchers for medical expert system design. Each of them utilizes some form of machine learning & deep learning interfaces. These interfaces assist in high-efficiency feature extraction, and disease classification via augmentation & interpolation of input datasets. But the scalability & performance of these models is limited due to their general-purpose architectural characteristics. Models that have high scalability, showcase moderate accuracy & precision performance when tested under multiple types of medical datasets. To improve this scalability, while maintaining high accuracy, precision & recall performance, a novel intrinsically scalable model for medical disease identification using hybrid swarm intelligence with augmented transfer learning is proposed in this text. The proposed ISM2SAITL model initially converts all input datasets into dynamically scaled & quantized 2D representations, which assists in deploying effective feature extraction methods.

A Method for the Study of Cerebellar Cognitive Function-Re-Cognition and Validation of Error-Related Potentials.

The cerebellar region has four times as many brain cells as the brain, but whether the cerebellum functions in cognition, and how it does so, remain unexplored. In order to verify whether the cerebellum is involved in cognition, we chose to investigate whether the cerebellum is involved in the process of error judgment. We designed an experiment in which we could activate the subject’s error-related potentials (ErrP). We recruited 26 subjects and asked them to wear EEG caps with cerebellar regions designed by us to participate in the experiment so that we could record their EEG activity throughout the experiment. We successfully mitigated the majority of noise interference after a series of pre-processing of the data collected from each subject. Our analysis of the preprocessed data revealed that our experiment successfully activated ErrP, and that the EEG signals, including the cerebellum, were significantly different when subjects made errors compared to when they made correct judgments. We designed a feature extraction method that requires selecting channels with large differences under different classifications, firstly by extracting the time-frequency features of these channels, and then screening these features with sequence backward feature (SBS) selection. We use the extracted features as the input and different event types in EEG data as the labels for multiple classifiers to classify the data in the executive and feedback segments, where the average accuracy for two-class classification of executive segments can reach 80.5%. The major contribution of our study is the discovery of the presence of ErrP in cerebellar regions and the extraction of an effective feature extraction method for EEG data.

Feature Extraction for Evaluating the Complexity of Electromagnetic Environment Based on Adaptive Multiscale Morphological Gradient and Nonnegative Matrix Factorization

In the current battlefield space, with the massive application of electromagnetic equipment, the electromagnetic environment in the battlefield space tends to be complex, which can lead to the electromagnetic equipment and personnel in the battlefield space receiving interference from the electromagnetic environment signal. To protect the safety of personnel and equipment quality, it is necessary to evaluate the complexity of the electromagnetic environment signal research, to use the corresponding measures. However, there is still little research related to the evaluation of the complexity of electromagnetic environmental signals. In this paper, a feature extraction method for electromagnetic environmental signals based on adaptive multiscale morphological gradient filtering and a nonnegative matrix factorization algorithm is proposed. First, the electromagnetic environment signal is filtered by AMMG, and then the filtered signal is processed by NMF for feature extraction. Finally, the complex electromagnetic environment signals after feature extraction are evaluated and classified by the SVM method. The results show that the evaluation results have good classification accuracy, and this paper provides an effective feature extraction method for the complexity of electromagnetic environment signals.

Future activity prediction of multiple sclerosis with 3D MRI using 3D discrete wavelet transform

• The future activity of Multiple Sclerosis is predicted using 3D MRI data. • 3D-DWT is used as a feature extraction method from 3D MRI data. • The relationships between MRI slices preserves by studying with 3D MRI. • The spatial and spectral location features are extracted by 3D-DWT. • Ten different wavelet families have been analyzed using machine learning algorithms. Multiple Sclerosis (MS) is a chronic and autoimmune neurological disease that is frequently seen especially in young people. MS lesions that can be seen with magnetic resonance imaging (MRI) findings are important biomarkers that provide information about the clinical prognosis and activity of the disease. The presence of new MS lesions is associated with future disease activity. This study aims to predict the future activity of MS using the 3D discrete wavelet transform (DWT) as a feature extraction method from 3D MRI. The 3D-DWT can be used as it provides spatial and spectral location features of MS lesions without losing their relationship between MRI slices. Ten different wavelet families of DWT are used individually, each of them is classified by six machine learning algorithms, and their feature extraction performances are compared. The highest F1-score, Precision, and Recall of 95.0% are obtained by the support vector machine algorithm on the SYM4, SYM8, and Haar wavelet families in the 3D MRI dataset consisting of 40 patients based on 5-fold cross validation. The results show that the 3D-DWT method is an effective method for feature extraction in predicting the future activity of MS.

A novel part-level feature extraction method for fine-grained vehicle recognition

In this paper, we propose a novel part-level feature extraction method to enhance the discriminative ability of deep convolutional features for the task of fine-grained vehicle recognition. Generally, the challenges for fine-grained vehicle recognition are mainly caused by the subtle visual differences between part regions of vehicles. Therefore, it is essential to extract discriminative features from part regions. Many existing methods, especially deep convolutional neural networks (D-CNNs), tend to detect the discriminative part regions explicitly or learn the part information implicitly through network restructuring and neglect the abundant part-level information contained in the high-level features generated by CNNs. In light of this, we propose a simple and effective part-level feature extraction method to enhance the representation of part-level features within the global features of target object generated by the backbone networks. The proposed method is built on the deep convolutional layers from which the discriminative part features could be integrated and extracted accordingly. More specifically, a basic feature grouping module is adopted to integrate the feature maps of deep convolutional layers into groups in each of which the related discriminative parts are assembled. The feature grouping process is performed in a multi-stage manner to ensure the integration process. Then a fusion module follows to model the coarse-to-fine relationship of the part features and further ensure the integrity and effectiveness of the part features. We conduct comparison experiments on public datasets, and the results show that the proposed method achieves comparable performance with state-of-the-art algorithms.

A temporal clustering method fusing deep convolutional autoencoders and dimensionality reduction methods and its application in air quality visualization

With the development of technologies, big data has become a common feature in all fields, and many of them are time-series data. To find out the data regular over time, effective feature extraction methods and visualization methods are required. In this paper, we firstly use the deep convolutional auto-encoder method to extract the features in the specified time range. Then, the dimension of the extracted features is reduced using different dimension reduction methods, and the results are presented by 2D projection. The results show the distributions of different time-series patterns. Air quality data of Hong Kong are used to verify the effectiveness of the method, and the method is also compared with different dimensionality reduction methods directly, and the proposed method in our paper shows great advantages in time clustering. • Feature extraction methods and visualization method are required for big data. • An improved DCAE based dimension reduction visualization method is proposed. • The DCAE based method is used to reduce the data dimension. • Several dimension reduction methods are used to realize projection visualization. • The air quality data of Hong Kong are used to test the algorithm.

Overall optimization of CSP based on ensemble learning for motor imagery EEG decoding

• A new algorithm framework based on ensemble learning is proposed for the overall optimization of CSP. • A new temporal-spatial-frequency feature joint optimization method is proposed. • A new method of generating feature diversity is proposed. • A large number of data sets are used to verify the effectiveness, universality, and robustness of the proposed method. The common spatial pattern (CSP) is an effective feature extraction method in motor imagery-based brain-computer interface (BCI) system. However, CSP also has many defects. Existing CSP improvement methods only make partial improvements, without considering the overall optimization of CSP. In this paper, a new ensemble learning algorithm framework is proposed to improve the decoding performance of CSP, in which the regularization, temporal-spatial-frequency joint optimization, and pair number of spatial filters for CSP are comprehensively considered. First, a new temporal-spatial-frequency feature extraction method based on Tikhonov regularization CSP (TRCSP) is proposed, multiple feature subsets with diversity are extracted by TRCSP with different time windows, regularization parameters, and pair numbers of spatial filters. Second, the least absolute shrinkage and selection operator (LASSO) as base classification model is used for feature selection and classification, in which multiple diversified base classification models are trained. Finally, the base classification models with diversity and higher accuracy are used for ensemble model construction using a new integration rule, during which most of the temporal-spatial-frequency information is fully excavated and utilized. The effectiveness of the proposed method is verified by five motor imagery data sets and the average classification accuracy of all data sets is 85.99%. Compared with the existing CSP methods, the proposed method achieved a better classification effect, and with a small amount of calculation, low model complexity, and high robustness.

An effective feature extraction method via spectral-spatial filter discrimination analysis for hyperspectral image

An effective feature extraction method via spectral-spatial filter discrimination analysis for hyperspectral image

Research on Dance Movement Recognition Based on Multi-Source Information

A huge number of scientific research institutions and scholars are now researching this topic in depth, with promising results. Meanwhile, research development in dance visual frequency movement detection is rather modest due to the high complexity of dance movement and the challenges of human body self-shielding in dance performance. Aiming at the problem of the combination of motion recognition and dance video, the feature extraction, representation, and motion recognition methods based on dance video are emphatically studied. This paper studies an effective feature extraction method according to the characteristics of dance movements. Firstly, each dance movement video in the data set is separated into equal sections, and the edge characteristics of all video pictures in each segment are gathered into one image, from which the direction gradient histogram features are extracted. Secondly, a group of directional gradient histogram feature vectors is used to represent the local appearance information and shape features of the video dance moves. In view of the existing problem of heterogeneous feature fusion, this paper chooses the multi-core learning method to fuse the three kinds of features for dance movement recognition. Finally, the effectiveness of the proposed dance movement detection algorithm is tested using the Dance DB data set from the University of Cyprus and the Folk Dance data set from my laboratory. Experimental results show that the proposed algorithm can maintain a certain recognition rate for relatively complex dance movements and can still ensure a certain accuracy when the background and target are easily confused. This also confirms the efficacy of the movement recognition system used in this paper for recognizing dance movements.

Optical signal monitoring using multi-task residual network

Optical signal monitoring, including optical performance monitoring (OPM) and bit rate and modulation format identification (BR-MFI), has attracted considerable interest owing to the advantages in automatic system diagnosis and the establishment of adaptive optical networks. However, the lack of effective multi-parameter shared feature extraction methods makes the optical signal monitoring still elusive and hinders its practical application. Exploiting shortcut connections that allow features to propagate across multiple layers to effectively reduce the low-level feature loss, we introduce a multi-task residual network (MT-ResNet) to extract optical signal feature information from eye diagrams. The results show that the model exhibits a multi-parameter shared feature extraction ability and OPM and BR-MFI can be directly realized. The average root-mean-square errors of optical signal-to-noise ratio, chromatic dispersion, and differential group delay monitoring are 0.84 dB, 1.8 ps/nm, and 0.18 ps, respectively. The accuracies of BR-MFI reach 100%. We demonstrate that the model is robust by investigating its feature extraction ability in the case of increased fiber nonlinearity and untrained samples, and the universality of the model is proved by applying it to a long-reach optical coherent communication system. Our simulation results indicate that this MT-ResNet can be deployed to analyze the signal quality cost-effectively, with a high potential for future dynamic and heterogeneous optical networks.