Papers by Ali Abdollahi Gharbali
Most of the existing methods for automatic sleep stage classification are relying on hand-crafted... more Most of the existing methods for automatic sleep stage classification are relying on hand-crafted features. In this paper, the goal is to develop a deep learning-based method that automatically exploits time-frequency spectrum of Electroencephalogram (EEG) signal, removing the need for manual feature extraction. Using Continuous Wavelet Transform (CWT), we extracted the time-frequency spectrogram for EEG signal of 10 healthy subjects and converted to RGB images. The images were classified using transfer learning of a pre-trained Convolutional Neural Network (CNN), AlexNet. The proposed method was evaluated using a publicly available dataset. Evaluation results show that our method can achieve state of the art accuracy, while having higher overall sensitivity.
Objective: In this paper, the contribution of distance-based features to automatic sleep stage cl... more Objective: In this paper, the contribution of distance-based features to automatic sleep stage classification is investigated. The potency of these features is analyzed individually and in combination with 48 conventionally used features.
Methods: The distance-based set consists of 32 features extracted by calculating Itakura, Itakura-Saito and COSH distances of autoregressive and spectral coefficients of Electrocardiography (EEG) (C3-A2), Left EOG, Chin EMG and ECG signals. All the evaluations are performed on three feature sets: distance-based, conventional and total (combined distance based and conventional). Six ranking methods were used to find the top features with the highest discrimination ability in each set. The ranked feature lists were evaluated using k-Nearest Neighbor (kNN), Artificial Neural Network (ANN), and Decision-tree-based multi-SVM (DSVM) classifiers for five sleep stages including Wake, REM, N1, N2 and N3. Furthermore, the ability of distance-based and conventional features to discriminate between each pair of sleep stages was evaluated using t-test, a hypothesis testing method. Results: Distance-based features occupied 25% of top-ranked features. Simulation results showed that using distance-based features together with conventional features can lead to an enhancement of accuracy. The best classification accuracy (85.5%) was achieved by DSVM classifier and 13 features selected by mRMR-MID and normalized with Min-Max method for total feature set, where two of them were from the distance-based feature set. The t-test results show that distance-based features outperform conventional features in discriminating be- tween N1 and REM stages that is usually a challenge for classification systems.
Conclusion: Distance-based features have a positive contribution to sleep stage classification, including enhancement of accuracy and better REM-N1 discrimination ability.
Significance: The main motivation for this work was to evaluate new features to characterize each sleep stage in such a way that extracted features were more powerful than conventional features, to distinguish sleep stages from each other, and to improve classifiers accuracy.
Background
Nowadays, sleep quality is one of the most important measures of healthy life, especi... more Background
Nowadays, sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using polysomnographic (PSG) signals is the traditional way of assessing sleep quality. However, the manual process of sleep stage classification is time-consuming, subjective and costly. Therefore, in order to improve the accuracy and efficiency of the sleep stage classification, researchers have been trying to develop automatic classification algorithms. Automatic sleep stage classification mainly consists of three steps: pre-processing, feature extraction and classification. Since classification accuracy is deeply affected by the extracted features, a poor feature vector will adversely affect the classifier and eventually lead to low classification accuracy. Therefore, special attention should be given to the feature extraction and selection process.
Methods
In this paper the performance of seven feature selection methods, as well as two feature rank aggregation methods, were compared. Pz-Oz EEG, horizontal EOG and submental chin EMG recordings of 22 healthy males and females were used. A comprehensive feature set including 49 features was extracted from these recordings. The extracted features are among the most common and effective features used in sleep stage classification from temporal, spectral, entropy-based and nonlinear categories. The feature selection methods were evaluated and compared using three criteria: classification accuracy, stability, and similarity.
Results
Simulation results show that MRMR-MID achieves the highest classification performance while Fisher method provides the most stable ranking. In our simulations, the performance of the aggregation methods was in the average level, although they are known to generate more stable results and better accuracy.
Conclusions
The Borda and RRA rank aggregation methods could not outperform significantly the conventional feature ranking methods. Among conventional methods, some of them slightly performed better than others, although the choice of a suitable technique is dependent on the computational complexity and accuracy requirements of the user.
In this paper a deep learning based dimension reduction, feature transformation and classificatio... more In this paper a deep learning based dimension reduction, feature transformation and classification method is proposed for automatic sleep stage classification. In order to enhance the feature vector, before feeding it to the deep network, a discriminative feature selection method is applied for removing the features with minimum information. Two-layer Stacked Sparse Autoencoder together with Softmax classifier is selected as the deep network model. The performance of the proposed method is compared with Softmax and k-nearest neighbour classifiers. Simulation results show that proposed deep learning structure outperformed others in terms of classification accuracy.
ABSTRACT Transformer inrush currents are high magnitude, harmonic-rich currents generated when tr... more ABSTRACT Transformer inrush currents are high magnitude, harmonic-rich currents generated when transformer cores are driven into saturation during energization. In this paper an efficient method for detection of inrush current in distribution transformer based on wavelet transform is presented. Using this method inrush current can be discriminate from other transients such as capacitor switching, load switching and single phase to ground fault. Wavelet transform is used for decomposition of signals and Learning Vector Quantizer(LVQ) neural network used for classification. Inrush current data and other transients are obtained by simulation using EMTP program. Results show that the proposed procedure is efficient in identifying inrush current from other events.
Sleep quality is one of the most important measures of healthy life, especially considering the h... more Sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using multi-channel recordings like poly-somnographic (PSG) signals is an effective way of assessing sleep quality. However, manual sleep stage classification is time-consuming, tedious and highly subjective. To overcome this, automatic sleep classification was proposed, in which pre-processing, feature extraction and classification are the three main steps. Since the classification accuracy is deeply affected by the features selection, in this paper several feature selection methods as well as rank aggregation methods are compared. Feature selection methods are evaluated by three criteria: accuracy, stability and similarity. For classification two different classifiers (k-nearest neighbor and multilayer feedforward neural network) were utilized. Simulation results show that MRMR-MID achieves highest classification performance while Fisher method provides the most stable rankings.
In this paper, a new algorithm is proposed for artifact removing of sleep electroencephalogram (E... more In this paper, a new algorithm is proposed for artifact removing of sleep electroencephalogram (EEG) with application in sleep stage classification. Rather than other works which used artificial noise, in this study real EEG data contaminated with electro-oculogram (EOG) and electromyogram (EMG) are used for evaluating the proposed artifact removal algorithm's efficiency using classification accuracy. The artifact detection is performed by thresholding the EEG-EOG and EEG-EMG cross correlation coefficients. Then, the segments considered contaminated are denoised by normalized least-mean squares (NLMS) adaptive filtering technique. Using a single EEG channel, four sleep stages consisting of Awake, Stage1 + REM, Stage 2 and Slow Wave Stage (SWS) are classified. A wavelet packet (WP) based feature set together with artificial neural network (ANN) are deployed for sleep stage classification purpose. Simulation results show that artifact removed EEG allows a classification accuracy improvement of around 14%.
Transformer inrush currents are high magnitude,
harmonic-rich currents generated when transformer... more Transformer inrush currents are high magnitude,
harmonic-rich currents generated when transformer cores are
driven into saturation during energization. This paper presents
an S-Transform based Probabilistic Neural Network (PNN)
classifier for recognition of inrush current. Using this method
inrush current can be discriminate from other transients such as
capacitor switching, load switching and single phase to ground
fault. S-transform is used for feature extraction and PNN is used
for classification. Inrush current data and other transients are
obtained by simulation using EMTP program. The simulation
results reveal that the combination of S-Transform and PNN can
effectively detect inrush current from other events.
Detection of Inrush Current Based On Wavelet
Transform and LVQ Neural Network
G.
Combination approaches provide an interesting
way to improve adaptive filter performance. In this... more Combination approaches provide an interesting
way to improve adaptive filter performance. In this
paper, we use one of such approaches known as ‘convex
combination of two transversal filters’ for speech
enhancement. In addition, both fixed and variable step
size cases are examined. As our simulation results show,
the variable step size version of the algorithm
outperforms the other case in a sense of convergence
rate and steady-state error.
Uploads
Papers by Ali Abdollahi Gharbali
Methods: The distance-based set consists of 32 features extracted by calculating Itakura, Itakura-Saito and COSH distances of autoregressive and spectral coefficients of Electrocardiography (EEG) (C3-A2), Left EOG, Chin EMG and ECG signals. All the evaluations are performed on three feature sets: distance-based, conventional and total (combined distance based and conventional). Six ranking methods were used to find the top features with the highest discrimination ability in each set. The ranked feature lists were evaluated using k-Nearest Neighbor (kNN), Artificial Neural Network (ANN), and Decision-tree-based multi-SVM (DSVM) classifiers for five sleep stages including Wake, REM, N1, N2 and N3. Furthermore, the ability of distance-based and conventional features to discriminate between each pair of sleep stages was evaluated using t-test, a hypothesis testing method. Results: Distance-based features occupied 25% of top-ranked features. Simulation results showed that using distance-based features together with conventional features can lead to an enhancement of accuracy. The best classification accuracy (85.5%) was achieved by DSVM classifier and 13 features selected by mRMR-MID and normalized with Min-Max method for total feature set, where two of them were from the distance-based feature set. The t-test results show that distance-based features outperform conventional features in discriminating be- tween N1 and REM stages that is usually a challenge for classification systems.
Conclusion: Distance-based features have a positive contribution to sleep stage classification, including enhancement of accuracy and better REM-N1 discrimination ability.
Significance: The main motivation for this work was to evaluate new features to characterize each sleep stage in such a way that extracted features were more powerful than conventional features, to distinguish sleep stages from each other, and to improve classifiers accuracy.
Nowadays, sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using polysomnographic (PSG) signals is the traditional way of assessing sleep quality. However, the manual process of sleep stage classification is time-consuming, subjective and costly. Therefore, in order to improve the accuracy and efficiency of the sleep stage classification, researchers have been trying to develop automatic classification algorithms. Automatic sleep stage classification mainly consists of three steps: pre-processing, feature extraction and classification. Since classification accuracy is deeply affected by the extracted features, a poor feature vector will adversely affect the classifier and eventually lead to low classification accuracy. Therefore, special attention should be given to the feature extraction and selection process.
Methods
In this paper the performance of seven feature selection methods, as well as two feature rank aggregation methods, were compared. Pz-Oz EEG, horizontal EOG and submental chin EMG recordings of 22 healthy males and females were used. A comprehensive feature set including 49 features was extracted from these recordings. The extracted features are among the most common and effective features used in sleep stage classification from temporal, spectral, entropy-based and nonlinear categories. The feature selection methods were evaluated and compared using three criteria: classification accuracy, stability, and similarity.
Results
Simulation results show that MRMR-MID achieves the highest classification performance while Fisher method provides the most stable ranking. In our simulations, the performance of the aggregation methods was in the average level, although they are known to generate more stable results and better accuracy.
Conclusions
The Borda and RRA rank aggregation methods could not outperform significantly the conventional feature ranking methods. Among conventional methods, some of them slightly performed better than others, although the choice of a suitable technique is dependent on the computational complexity and accuracy requirements of the user.
harmonic-rich currents generated when transformer cores are
driven into saturation during energization. This paper presents
an S-Transform based Probabilistic Neural Network (PNN)
classifier for recognition of inrush current. Using this method
inrush current can be discriminate from other transients such as
capacitor switching, load switching and single phase to ground
fault. S-transform is used for feature extraction and PNN is used
for classification. Inrush current data and other transients are
obtained by simulation using EMTP program. The simulation
results reveal that the combination of S-Transform and PNN can
effectively detect inrush current from other events.
way to improve adaptive filter performance. In this
paper, we use one of such approaches known as ‘convex
combination of two transversal filters’ for speech
enhancement. In addition, both fixed and variable step
size cases are examined. As our simulation results show,
the variable step size version of the algorithm
outperforms the other case in a sense of convergence
rate and steady-state error.
Methods: The distance-based set consists of 32 features extracted by calculating Itakura, Itakura-Saito and COSH distances of autoregressive and spectral coefficients of Electrocardiography (EEG) (C3-A2), Left EOG, Chin EMG and ECG signals. All the evaluations are performed on three feature sets: distance-based, conventional and total (combined distance based and conventional). Six ranking methods were used to find the top features with the highest discrimination ability in each set. The ranked feature lists were evaluated using k-Nearest Neighbor (kNN), Artificial Neural Network (ANN), and Decision-tree-based multi-SVM (DSVM) classifiers for five sleep stages including Wake, REM, N1, N2 and N3. Furthermore, the ability of distance-based and conventional features to discriminate between each pair of sleep stages was evaluated using t-test, a hypothesis testing method. Results: Distance-based features occupied 25% of top-ranked features. Simulation results showed that using distance-based features together with conventional features can lead to an enhancement of accuracy. The best classification accuracy (85.5%) was achieved by DSVM classifier and 13 features selected by mRMR-MID and normalized with Min-Max method for total feature set, where two of them were from the distance-based feature set. The t-test results show that distance-based features outperform conventional features in discriminating be- tween N1 and REM stages that is usually a challenge for classification systems.
Conclusion: Distance-based features have a positive contribution to sleep stage classification, including enhancement of accuracy and better REM-N1 discrimination ability.
Significance: The main motivation for this work was to evaluate new features to characterize each sleep stage in such a way that extracted features were more powerful than conventional features, to distinguish sleep stages from each other, and to improve classifiers accuracy.
Nowadays, sleep quality is one of the most important measures of healthy life, especially considering the huge number of sleep-related disorders. Identifying sleep stages using polysomnographic (PSG) signals is the traditional way of assessing sleep quality. However, the manual process of sleep stage classification is time-consuming, subjective and costly. Therefore, in order to improve the accuracy and efficiency of the sleep stage classification, researchers have been trying to develop automatic classification algorithms. Automatic sleep stage classification mainly consists of three steps: pre-processing, feature extraction and classification. Since classification accuracy is deeply affected by the extracted features, a poor feature vector will adversely affect the classifier and eventually lead to low classification accuracy. Therefore, special attention should be given to the feature extraction and selection process.
Methods
In this paper the performance of seven feature selection methods, as well as two feature rank aggregation methods, were compared. Pz-Oz EEG, horizontal EOG and submental chin EMG recordings of 22 healthy males and females were used. A comprehensive feature set including 49 features was extracted from these recordings. The extracted features are among the most common and effective features used in sleep stage classification from temporal, spectral, entropy-based and nonlinear categories. The feature selection methods were evaluated and compared using three criteria: classification accuracy, stability, and similarity.
Results
Simulation results show that MRMR-MID achieves the highest classification performance while Fisher method provides the most stable ranking. In our simulations, the performance of the aggregation methods was in the average level, although they are known to generate more stable results and better accuracy.
Conclusions
The Borda and RRA rank aggregation methods could not outperform significantly the conventional feature ranking methods. Among conventional methods, some of them slightly performed better than others, although the choice of a suitable technique is dependent on the computational complexity and accuracy requirements of the user.
harmonic-rich currents generated when transformer cores are
driven into saturation during energization. This paper presents
an S-Transform based Probabilistic Neural Network (PNN)
classifier for recognition of inrush current. Using this method
inrush current can be discriminate from other transients such as
capacitor switching, load switching and single phase to ground
fault. S-transform is used for feature extraction and PNN is used
for classification. Inrush current data and other transients are
obtained by simulation using EMTP program. The simulation
results reveal that the combination of S-Transform and PNN can
effectively detect inrush current from other events.
way to improve adaptive filter performance. In this
paper, we use one of such approaches known as ‘convex
combination of two transversal filters’ for speech
enhancement. In addition, both fixed and variable step
size cases are examined. As our simulation results show,
the variable step size version of the algorithm
outperforms the other case in a sense of convergence
rate and steady-state error.