Set Of Extracted Features Research Articles

ScHiClassifier: a deep learning framework for cell type prediction by fusing multiple feature sets from single-cell Hi-C data.

Single-cell high-throughput chromosome conformation capture (Hi-C) technology enables capturing chromosomal spatial structure information at the cellular level. However, to effectively investigate changes in chromosomal structure across different cell types, there is a requisite for methods that can identify cell types utilizing single-cell Hi-C data. Current frameworks for cell type prediction based on single-cell Hi-C data are limited, often struggling with features interpretability and biological significance, and lacking convincing and robust classification performance validation. In this study, we propose four new feature sets based on the contact matrix with clear interpretability and biological significance. Furthermore, we develop a novel deep learning framework named scHiClassifier based on multi-head self-attention encoder, 1D convolution and feature fusion, which integrates information from these four feature sets to predict cell types accurately. Through comprehensive comparison experiments with benchmark frameworks on six datasets, we demonstrate the superior classification performance and the universality of the scHiClassifier framework. We further assess the robustness of scHiClassifier through data perturbation experiments and data dropout experiments. Moreover, we demonstrate that using all feature sets in the scHiClassifier framework yields optimal performance, supported by comparisons of different feature set combinations. The effectiveness and the superiority of the multiple feature set extraction are proven by comparison with four unsupervised dimensionality reduction methods. Additionally, we analyze the importance of different feature sets and chromosomes using the "SHapley Additive exPlanations" method. Furthermore, the accuracy and reliability of the scHiClassifier framework in cell classification for single-cell Hi-C data are supported through enrichment analysis. The source code of scHiClassifier is freely available at https://github.com/HaoWuLab-Bioinformatics/scHiClassifier.

A Study of Occluded Person Re-Identification for Shared Feature Fusion with Pose-Guided and Unsupervised Semantic Segmentation

The human body is often occluded by a variety of obstacles in the monitoring system, so occluded person re-identification is still a long-standing challenge. Recent methods based on pose guidance or external semantic clues have improved the representation and related performance of features; there are still problems, such as weak model representation and unreliable semantic clues. To solve the above problems, we proposed a feature extraction network, named shared feature fusion with pose-guided and unsupervised semantic segmentation (SFPUS). This network will extract more discriminative features and reduce the occlusion noise on pedestrian matching. Firstly, the multibranch joint feature extraction module (MFE) is used to extract feature sets containing pose information and high-order semantic information. This module not only provides robust extraction capabilities but can also precisely segment occlusion and the body. Secondly, in order to obtain multiscale discriminant features, the multiscale correlation feature matching fusion module (MCF) is used to match the two feature sets, and the Pose–Semantic Fusion Loss is designed to calculate the similarity of the feature sets between different modes and fuse them into a feature set. Thirdly, to solve the problem of image occlusion, we use unsupervised cascade clustering to better prevent occlusion interference. Finally, performances of the proposed method and various existing methods are compared on the Occluded-Duke, Occluded-ReID, Market-1501 and Duke-MTMC datasets. The accuracy of Rank-1 reached 65.7%, 80.8%, 94.8% and 89.6%, respectively, and the mAP accuracy reached 58.8%, 72.5%, 91.8% and 80.1%. The experiment results demonstrate that our proposed SFPUS holds promising prospects and performs admirably compared with state-of-the-art methods.

Novel interpretable Feature set extraction and classification for accurate atrial fibrillation detection from ECGs

ObjectiveWe present a novel method for detecting atrial fibrillation (AFib) by analyzing Lead II electrocardiograms (ECGs) using a unique set of features. MethodsFor this purpose, we used specific signal processing techniques, such as proper orthogonal decomposition, continuous wavelet transforms, discrete cosine transform, and standard cross-correlation, to extract 48 features from the ECGs. Thus, our approach aims to more effectively capture AFib signatures, such as beat-to-beat variability and fibrillatory waves, than traditional metrics. Moreover, our features were designed to be physiologically interpretable. Subsequently, we incorporated an XGBoost-based ECG classifier to train and evaluate it using the publicly available ‘Training’ dataset of the 2017 PhysioNet Challenge, which includes ‘Normal,’ ‘AFib,’ ‘Other,’ and ‘Noisy’ ECG categories. ResultsOur method achieved an accuracy of 96 % and an F1-score of 0.83 for ‘AFib’ detection and 80 % accuracy and 0.85 F1-score for ‘Normal’ ECGs. Finally, we compared our method's performance with the top-classifiers from the 2017 PhysioNet Challenge, namely ENCASE, Random Forest, and Cascaded Binary. As reported by Clifford et al., 2017, these three best performing models scored 0.80, 0.83, 0.82, in terms of F1-score for ‘AFib’ detection, respectively. ConclusionDespite using significantly fewer features than the competition's state-of-the-art ECG detection algorithms (48 vs. 150–622), our model achieved a comparable F1-score of 0.83 for successful ‘AFib’ detection. Significance: The interpretable features specifically designed for ‘AFib’ detection results in our method's adaptability in clinical settings for real-time arrhythmia detection and is even effective with short ECGs (<10 heartbeats).

Single channel speech enhancement using time-frequency attention mechanism based nested U-net model

Deep-learning models have used attention mechanisms to improve quality and intelligibility of noisy speech, demonstrating the effectiveness of attention mechanisms. We rely on either spatial or temporal-based attention mechanisms, resulting in severe information loss. In this paper, a time-frequency attention mechanism with a nested U-network (TFANUNet) is proposed for single-channel speech enhancement. By using TFA, learns which channel, frequency and time information is more significant for speech enhancement. Basically, the proposed model is an encoder-decoder model, where each layer in the encoder and decoder is followed by a nested dense residual dilated DensNet (NDRD) based multi-scale context aggression block. NDRD involves multiple dilated convolutions with different dilatation factors to explore the large receptive area at different scales simultaneously. NDRD avoids the aliasing problem in DenseNet. We integrated the TFA and NDRD blocks into the proposed model to enable refined feature set extraction without information loss and utterance-level context aggregation, respectively. Under seen and unseen noise conditions, the proposed TFAD3MNet model produces an average of 87.02% and 85.04% of STOI values, and 3.19 and 3.01 averaged PESQ values. The trainable parameters of proposed model are 2.09 million,which is very less compared to baselines. TFANUNet model results outperform baselines in terms of STOI and PESQ.

Independent Video Steganalysis Framework Perspective of Secure Video Processing

We have designed an algorithm of cross-domain feature set extraction for cross-domain video steganalysis of video steganography in multiple domains. For video steganography, we implemented the recently proposed steganography methods such as PMs, MVs, and IPMs. The outcomes of these techniques have fed as input to the proposed Steganalysis approach. We have designed a cross-domain technique for Steganalysis in which the global feature set is extracted according to common statistical properties. After the extraction of global features, the domain-specific features have been extracted in the local features set. Both global and local features are set to form the cross-domain steganalysis technique. The classification has been performed by using the conventional machine learning classifiers such as Support Vector Machine (SVM) and Artificial Neural Network (ANN).

EEG channel selection using Gramian Angular Fields and spectrograms for energy data visualization

Electroencephalography (EEG)-based brain-computer interfaces (BCIs) have a wide range of applications in affect recognition. The usage of irrelevant information channels when decoding brain activity from different regions can negatively impact the task at hand. Therefore, further research is needed to determine the ideal channels for detecting superior performances. In this study, deep learning models and 2-D image representations are used to assess the efficacy of EEG channels for subjective valence responses to energy data visualizations. The EEG signals are converted into spectrograms and Gramian Angular Field (GAF) images. A hybrid Convolution Neural Network (CNN)-Long-Short Term Memory (LSTM) model and LSTM network are used to extract feature sets from the converted images. These feature sets, after reducing their dimensionality using a principal component analysis (PCA) method, are fed into a boosting classifier (AdaBoost). The performance metrics of both models and 2-D representations are compared. The LSTM and CNN-LSTM models with GAFs and spectrograms achieve state-of-the-art accuracies. The CNN-LSTM model achieves the highest performance when using spectrograms for the F8 channel, while the GAF method performs relatively lower for the F3 channel. The CNN-LSTM model with the spectrogram method produces reliable results, and hence, this method is used for further analysis of the remaining channel pairs. This study suggests that using a single EEG channel is more effective than using multiple channels for recognizing emotions in energy data visualizations. The proposed methodology is an efficient way to select channels for this purpose.

Decoding Autism: Uncovering patterns in brain connectivity through sparsity analysis with rs-fMRI data

Background:In the realm of neuro-disorders, precise diagnosis and treatment rely heavily on objective imaging-based biomarker identification. This study employs a sparsity approach on resting-state fMRI to discern relevant brain region connectivity for predicting Autism. New method:The proposed methodology involves four key steps: (1) Utilizing three probabilistic brain atlases to extract functionally homogeneous brain regions from fMRI data. (2) Employing a hybrid approach of Graphical Lasso and Akaike Information Criteria to optimize sparse inverse covariance matrices for representing the brain functional connectivity. (3) Employing statistical techniques to scrutinize functional brain structures in Autism and Control subjects. (4) Implementing both autoencoder-based feature extraction and entire feature-based approach coupled with AI-based learning classifiers to predict Autism. Results:The ensemble classifier with the extracted feature set achieves a classification accuracy of 84.7% ± 0.3% using the MSDL atlas. Meanwhile, the 1D-CNN model, employing all features, exhibits superior classification accuracy of 88.6% ± 1.7% with the Smith 2009 (rsn70) atlas. Comparison with existing method (s):The proposed methodology outperforms the conventional correlation-based functional connectivity approach with a notably high prediction accuracy of more than 88%, whereas considering all direct and noisy indirect region-based functional connectivity, the traditional methods bound the prediction accuracy within 70% to 79%. Conclusions:This study underscores the potential of sparsity-based FC analysis using rs-fMRI data as a prognostic biomarker for detecting Autism.

Deep ensemble of classifier for intrusion detection in WSN and improved attack mitigation process

SummaryAttacks have a significant negative impact on the wireless sensor network's network operations. They target to weaken the network layer, if they are not completely eradicated, the networks' ability to execute their desired function becomes collapsed. While intended to monitor such unexpected attacks, anomaly‐based intrusion detection systems (AIDS) have a high risk of false positives. This article intends to propose a Deep Ensemble Intrusion detection with a Self‐adaptive Sewing Training‐Based Optimization (DIED‐SASTO) model that includes the following steps: (a) preprocessing, (b) feature extraction, (c) optimal feature selection, and (d) detection. Initially, an improved class imbalance processing will take place to solve the imbalance problem during the preprocessing step. Subsequently, the features are extracted including higher‐order statistical features, improved entropy, and correlation‐based features during the feature extraction step. From the extracted feature set, as the curse of dimensionality becomes a serious challenge, it is significant to choose the optimal features. In the optimal feature selection process, a new Self‐adaptive Sewing Training‐Based Optimization (SASTO) is used. In addition, the intrusion detection will be employed based on the selected optimal features and trained using deep ensemble models like Deep maxout, Deep Belief Network (DBN), and Bidirectional Long short‐term memory (Bi‐LSTM). Once the presence of an attack is detected, it is mandatory to mitigate the attacker node form the network. For Attack Mitigation, improved entropy will be progressed. The effectiveness of the DIED‐SASTO is evaluated over conventional methods in terms of various metrics.

Optimized Convolutional Forest by Particle Swarm Optimizer for Pothole Detection

Poor road maintenance leads to potholes on the road. Potholes are responsible for road accidents and even deaths in developed and developing countries. Detecting and filling road potholes is an essential part of road maintenance. Sustaining a reliable and safe road for communication depends on pothole detection. This study presents a novel combination of a convolutional neural network and an optimized machine-learning model by a heuristic algorithm for pothole detection. The proposed method comprises a shallow convolutional neural network for feature extraction and an optimized random forest model for pothole detection. The proposed model initially uses the shallow convolutional layer to extract feature sets from input pictures. Then, the particle swarm optimizer is used to eliminate irrelevant features. Finally, a combination of random forest and a particle swarm optimizer is used for pothole detection. Particle swarm optimization indicates the best subset of the extracted feature set for final pothole detection. We added 171 pictures to the already available 665 pothole pictures to evaluate the proposed method. The test set was isolated from the training set, and we trained the model on k-fold cross-validation. The experimental result indicates 99.37% accuracy, 99.37% precision, 99.38% sensitivity, and 99.38% F1-score for discriminating potholes from roads without potholes by proposed methods. The response time of the proposed method for pothole detection is 0.02 s. The proposed method can be utilized for real-time pothole detection.

Motion vector‐domain video steganalysis exploiting skipped macroblocks

AbstractVideo steganography has the potential to be used to convey illegal information, and video steganalysis is a vital tool to detect the presence of this illicit act. Currently, all the motion vector (MV)‐based video steganalysis algorithms extract feature sets directly from the MVs, but ignoring the embedding operation may perturb the statistical distribution of other video encoding elements, such as the skipped macroblocks (no direct MVs). This paper proposes a novel 11‐dimensional feature set to detect MV‐based video steganography based on the above observation. The proposed feature is extracted based on the skipped macroblocks by recompression calibration. Specifically, the feature consists of two components. The first is the probability distribution of motion vector prediction (MVP) difference, and the second is the probability distribution of partition state transfer. Extensive experiments on different conditions demonstrate that the proposed feature set achieves good detection accuracy, especially in lower embedding capacities. In addition, the loss of detection performance caused by recompression calibration using mismatched quantization parameters (QP) is within the acceptable range, so the proposed method can be used in practical scenarios.

An attention 3DUNET and visual geometry group-19 based deep neural network for brain tumor segmentation and classification from MRI

There has been an abrupt increase in brain tumor (BT) related medical cases during the past ten years. The tenth most typical type of tumor affecting millions of people is the BT. The cure rate can, however, rise if it is found early. When evaluating BT diagnosis and treatment options, MRI is a crucial tool. However, segmenting the tumors from magnetic resonance (MR) images is complex. The advancement of deep learning (DL) has led to the development of numerous automatic segmentation and classification approaches. However, most need improvement since they are limited to 2D images. So, this article proposes a novel and optimal DL system for segmenting and classifying the BTs from 3D brain MR images. Preprocessing, segmentation, feature extraction, feature selection, and tumor classification are the main phases of the proposed work. Preprocessing, such as noise removal, is performed on the collected brain MR images using bilateral filtering. The tumor segmentation uses spatial and channel attention-based three-dimensional u-shaped network (SC3DUNet) to segment the tumor lesions from the preprocessed data. After that, the feature extraction is done based on dilated convolution-based visual geometry group-19 (DCVGG-19), making the classification task more manageable. The optimal features are selected from the extracted feature sets using diagonal linear uniform and tangent flight included butterfly optimization algorithm. Finally, the proposed system applies an optimal hyperparameters-based deep neural network to classify the tumor classes. The experiments conducted on the BraTS2020 dataset show that the suggested method can segment tumors and categorize them more accurately than the existing state-of-the-art mechanisms. Communicated by Ramaswamy H. Sarma

HFRAS: design of a high-density feature representation model for effective augmentation of satellite images

AbstractEfficiently extracting features from satellite images is crucial for classification and post-processing activities. Many feature representation models have been created for this purpose. However, most of them either increase computational complexity or decrease classification efficiency. The proposed model in this paper initially collects a set of available satellite images and represents them via a hybrid of long short-term memory (LSTM) and gated recurrent unit (GRU) features. These features are processed via an iterative genetic algorithm, identifying optimal augmentation methods for the extracted feature sets. To analyse the efficiency of this optimization process, we model an iterative fitness function that assists in incrementally improving the classification process. The fitness function uses an accuracy &amp; precision-based feedback mechanism, which helps in tuning the hyperparameters of the proposed LSTM &amp; GRU feature extraction process. The suggested model used 100 k images, 60% allocated for training and 20% each designated for validation and testing purposes. The proposed model can increase classification precision by 16.1% and accuracy by 17.1% compared to conventional augmentation strategies. The model also showcased incremental accuracy enhancements for an increasing number of training image sets.

A novel multimodal framework for early diagnosis and classification of COPD based on CT scan images and multivariate pulmonary respiratory diseases

Background and objectiveChronic Obstructive Pulmonary Disease (COPD) is one of the world's worst diseases; its early diagnosis using existing methods like statistical machine learning techniques, medical diagnostic tools, conventional medical procedures, and other methods is challenging due to misclassification results of COPD diagnosis and takes a long time to perform accurate prediction. Due to the severe consequences of COPD, detection and accurate diagnosis of COPD at an early stage is essential. This paper aims to design and develop a multimodal framework for early diagnosis and accurate prediction of COPD patients based on prepared Computerized Tomography (CT) scan images and lung sound/cough (audio) samples using machine learning techniques, which are presented in this study. MethodThe proposed multimodal framework extracts texture, histogram intensity, chroma, Mel-Frequency Cepstral Coefficients (MFCCs), and Gaussian scale space from the prepared CT images and lung sound/cough samples. Accurate data from All India Institute Medical Sciences (AIIMS), Raipur, India, and the open respiratory CT images and lung sound/cough (audio) sample dataset validate the proposed framework. The discriminatory features are selected from the extracted feature sets using unsupervised ML techniques, and customized ensemble learning techniques are applied to perform early classification and assess the severity levels of COPD patients. ResultsThe proposed framework provided 97.50%, 98%, and 95.30% accuracy for early diagnosis of COPD patients based on the fusion technique, CT diagnostic model, and cough sample model. ConclusionFinally, we compare the performance of the proposed framework with existing methods, current approaches, and conventional benchmark techniques for early diagnosis.

PDHF: Effective phishing detection model combining optimal artificial and automatic deep features

Currently, the increasing number of high-volume phishing attacks is among the largest threats to networking environments on a daily basis. During such a severe attack, researchers prefer to extract numerous features to improve the accuracy of phishing detection. However, the redundant features that may exist in the extracted feature set may be adapted by phishing attackers, not only degrading detection performance but also shortening the effective time of the constructed detection models. To address these problems, this study proposes phishing detection based on hybrid features (PDHF), a novel phishing detection model based on a combination of optimal artificial and automatic deep learning features. The optimal artificial phishing features are obtained by removing redundant features based on the newly designed feature importance evaluation index and an improved bidirectional search algorithm. To extend the effective time of phishing detection, deep features are learned from URLs using a one-dimensional character convolutional neural network (CNN) and a disorderly quantized attention mechanism. The experimental results show that PDHF outperforms many state-of-the-art methods and achieves an accuracy of 0.9965, precision of 0.9942, recall of 0.9940, and F1-score of 0.9941. These results can help in the development of a security plug-in for clients, browsers, and various instant messaging tools that run on network edges, personal computers, smartphones, and other personal terminals.

Optimized Deep Learning Model for Flood Detection Using Satellite Images

The increasing amount of rain produces a number of issues in Kerala, particularly in urban regions where the drainage system is frequently unable to handle a significant amount of water in such a short duration. Meanwhile, standard flood detection results are inaccurate for complex phenomena and cannot handle enormous quantities of data. In order to overcome those drawbacks and enhance the outcomes of conventional flood detection models, deep learning techniques are extensively used in flood control. Therefore, a novel deep hybrid model for flood prediction (DHMFP) with a combined Harris hawks shuffled shepherd optimization (CHHSSO)-based training algorithm is introduced for flood prediction. Initially, the input satellite image is preprocessed by the median filtering method. Then the preprocessed image is segmented using the cubic chaotic map weighted based k-means clustering algorithm. After that, based on the segmented image, features like difference vegetation index (DVI), normalized difference vegetation index (NDVI), modified transformed vegetation index (MTVI), green vegetation index (GVI), and soil adjusted vegetation index (SAVI) are extracted. The features are subjected to a hybrid model for predicting floods based on the extracted feature set. The hybrid model includes models like CNN (convolutional neural network) and deep ResNet classifiers. Also, to enhance the prediction performance, the CNN and deep ResNet models are fine-tuned by selecting the optimal weights by the combined Harris hawks shuffled shepherd optimization (CHHSSO) algorithm during the training process. This hybrid approach decreases the number of errors while improving the efficacy of deep neural networks with additional neural layers. From the result study, it clearly shows that the proposed work has obtained sensitivity (93.48%), specificity (98.29%), accuracy (94.98%), false negative rate (0.02%), and false positive rate (0.02%) on analysis. Furthermore, the proposed DHMFP–CHHSSO displays better performances in terms of sensitivity (0.932), specificity (0.977), accuracy (0.952), false negative rate (0.0858), and false positive rate (0.036), respectively.

DiabPrednet: development of attention-based long short-term memory-based diabetes prediction model with optimal weighted feature fusion mechanism

ABSTRACT Machine learning is a computer technique that automatically learns from experience and enhances the effectiveness of producing more precise diabetes predictions. However, large, inclusive, high-quality datasets are needed for training the machine learning networks. In this research work, attention-based approaches are designed for predicting diabetes in the affected individuals. Initially, the collected diabetes data is given into the data cleaning to get noise-free data for the prediction task. Here, extracted feature set 1 is extracted from the Auto encoder, and extracted feature set 2 is extracted from the 1-Dimensional Convolutional Neural Network (1D-CNN). These two sets of extracted features are fused in the adaptive way that is weighted feature fusion. Here, the weight of the selected features is optimized by an Enhanced Path Finder Algorithm (EPFA) to get more accurate results. The weighted fused features are employed for the diabetes prediction phase, in which the developed Attention-based Long Short Term Memory (ALSTM) with architecture optimization by improved PFA for predicting diabetes in affected one. Throughout the result analysis, the designed method attains 95% accuracy and 92%precision rate. Finally, the analysis is made by the proposed and existing prediction methods to showcase the effective performance.

Nonadditive Entropy Application to Detrended Force Sensor Data to Indicate Balance Disorder of Patients with Vestibular System Dysfunction.

The healthy function of the vestibular system (VS) is of vital importance for individuals to carry out their daily activities independently and safely. This study carries out Tsallis entropy (TE)-based analysis on insole force sensor data in order to extract features to differentiate between healthy and VS-diseased individuals. Using a specifically developed algorithm, we detrend the acquired data to examine the fluctuation around the trend curve in order to consider the individual's walking habit and thus increase the accuracy in diagnosis. It is observed that the TE value increases for diseased people as an indicator of the problem of maintaining balance. As one of the main contributions of this study, in contrast to studies in the literature that focus on gait dynamics requiring extensive walking time, we directly process the instantaneous pressure values, enabling a significant reduction in the data acquisition period. The extracted feature set is then inputted into fundamental classification algorithms, with support vector machine (SVM) demonstrating the highest performance, achieving an average accuracy of 95%. This study constitutes a significant step in a larger project aiming to identify the specific VS disease together with its stage. The performance achieved in this study provides a strong motivation to further explore this topic.

A Comparative of Two-Dimensional Statistical Moment Invariants Features in Formulating an Automated Probabilistic Machine Learning Identification Algorithm for Forensic Application

IBIS, ALIS, EVOFINDER, and CONDOR are the massive ballistics computerised technological machines that have typically been utilised in forensic laboratories to automatically locate similarities between images of cartridge cases and bullets. However, it imposed a long execution time and requires physical interpretation to consolidate the analysis results when employing these market-available technologies to accomplish ballistics matching tasks. Therefore, the principal objective of this study is to propose an improvised automated probabilistic machine learning identification algorithm by extracting the two-dimensional (2D) statistical moment invariants from the segmented region of interest (ROI) corresponding to the cartridge case and bullets images. To pursue this principal objective, several 2D statistical moment invariants have been compared and tested to determine the most suitable feature set applied in the proposed identification algorithm. The 2D statistical moment invariants employed include Orthogonal Legendre moments (OLM), Hu moments (HM), Tsirikolias-Mertzois moments (TMM), Pan-Keane moments (PKM), and Central Geometric moments (CGM). Moreover, the proposed identification algorithm is also tested in different scenarios, including based on the classification of strength association measurements between the extracted feature sets. The empirical results in this article revealed that the proposed identification algorithm applied with the CGM comprising the weak association classification yielded the best identification accuracy rates, which are &gt;96.5% across all the sample sizes of the training set. These empirical results also conveyed that the superior proposed identification algorithm in this research could be developed as a mobile application for ballistics identification that can significantly reduce the time taken and conveniently perform the ballistics identification tasks.

Hybrid Feature-Learning-Based PSO-PCA Feature Engineering Approach for Blood Cancer Classification.

Acute lymphoblastic leukemia (ALL) is a lethal blood cancer that is characterized by an abnormal increased number of immature lymphocytes in the blood or bone marrow. For effective treatment of ALL, early assessment of the disease is essential. Manual examination of stained blood smear images is current practice for initially screening ALL. This practice is time-consuming and error-prone. In order to effectively diagnose ALL, numerous deep-learning-based computer vision systems have been developed for detecting ALL in blood peripheral images (BPIs). Such systems extract a huge number of image features and use them to perform the classification task. The extracted features may contain irrelevant or redundant features that could reduce classification accuracy and increase the running time of the classifier. Feature selection is considered an effective tool to mitigate the curse of the dimensionality problem and alleviate its corresponding shortcomings. One of the most effective dimensionality-reduction tools is principal component analysis (PCA), which maps input features into an orthogonal space and extracts the features that convey the highest variability from the data. Other feature selection approaches utilize evolutionary computation (EC) to search the feature space and localize optimal features. To profit from both feature selection approaches in improving the classification performance of ALL, in this study, a new hybrid deep-learning-based feature engineering approach is proposed. The introduced approach integrates the powerful capability of PCA and particle swarm optimization (PSO) approaches in selecting informative features from BPI mages with the power of pre-trained CNNs of feature extraction. Image features are first extracted through the feature-transfer capability of the GoogleNet convolutional neural network (CNN). PCA is utilized to generate a feature set of the principal components that covers 95% of the variability in the data. In parallel, bio-inspired particle swarm optimization is used to search for the optimal image features. The PCA and PSO-derived feature sets are then integrated to develop a hybrid set of features that are then used to train a Bayesian-based optimized support vector machine (SVM) and subspace discriminant ensemble-learning (SDEL) classifiers. The obtained results show improved classification performance for the ML classifiers trained by the proposed hybrid feature set over the original PCA, PSO, and all extracted feature sets for ALL multi-class classification. The Bayesian-optimized SVM trained with the proposed hybrid PCA-PSO feature set achieves the highest classification accuracy of 97.4%. The classification performance of the proposed feature engineering approach competes with the state of the art.

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.