Performance Improvement of Driving Fatigue Identification Based on Power Spectra and Connectivity Using Feature Level and Decision Level Fusions

Harvy, Jonathan; Sigalas, Evangelos; Thakor, Nitish V.; Bezerianos, Anastasios; Li, Junhua

doi:10.1109/embc.2018.8512259

Cited by 12 publications

(14 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The results show that classification accuracies were increased quickly to a local peak and then slightly increased to a balanced level for most cases. Based on the current study, the SPE reduced classification accuracy, which was different from our previous results [52], indicating that the same method has different performance on the different classification tasks and different datasets. It is worth noting that the FT was better than WPD based on the results of this study, which is not in agreement with the findings in other studies.…”

Section: Discussioncontrasting

confidence: 99%

EEG-Based Multiclass Workload Identification Using Feature Fusion and Selection

Pei

Wang

Bezerianos

et al. 2021

IEEE Trans. Instrum. Meas.

Self Cite

View full text Add to dashboard Cite

The effectiveness of workload identification is one of the critical aspects in a monitoring instrument of mental state. In this field, the workload is usually recognised as binary classes. There are scarce studies towards multi-class workload identification because the challenge of the success of workload identification is much tough, even though one more workload class is added. Besides, most of the existing studies only utilized spectral power features from individual channels but ignoring abundant inter-channel features that represent the interactions between brain regions. In this study, we utilized features representing intra-channel information and inter-channel information to classify multiple classes of workload based on EEG. We comprehensively compared each category of features contributing to workload identification and elucidated the roles of feature fusion and feature selection for the workload identification. The results demonstrated that feature combination (83.12% in terms of accuracy) enhanced the classification performance compared to individual feature categories (i.e., band power features, 75.90%; connection features, 81.72%, in terms of accuracy). With the Fscore feature selection, the classification accuracy was further increased to 83.47%. When the features of graph metric were fused, the accuracy was reached to 84.34%. Our study provided comprehensive performance comparisons between methods and feature categories for the multi-class workload identification and demonstrated that feature selection and fusion played an important role in the enhancement of workload identification. These results could facilitate further studies of multi-class workload identification and practical application of workload identification.

show abstract

Section: Discussioncontrasting

confidence: 99%

EEG-Based Multiclass Workload Identification Using Feature Fusion and Selection

Pei

Wang

Bezerianos

et al. 2021

IEEE Trans. Instrum. Meas.

Self Cite

View full text Add to dashboard Cite

show abstract

“…Most recently, high-order functional connectivity in both static and dynamic representations was found to have complementary information to low-order functional connectivity in fatigue detection [43]. These diverse kinds of features can be fused to improve fatigue classification [44]. In addition, transfer learning can be integrated with the proposed framework to enhance robustness of the performance across subjects and sessions [45].…”

Section: Ieee Transactions On Cognitive and Developmental Systemsmentioning

confidence: 99%

Regression-Based Continuous Driving Fatigue Estimation: Toward Practical Implementation

Bose

Wang

Dragomir

et al. 2020

IEEE Trans. Cogn. Dev. Syst.

Self Cite

View full text Add to dashboard Cite

Mental fatigue in drivers is one of the leading causes that give rise to traffic accidents. Electroencephalography (EEG) based driving fatigue studies showed promising performance in fatigue monitoring. However, complex methodologies are not suitable for practical implementation. In our simulation based setup that retained the constraints of real driving, we took a step closer to fatigue estimation in a practical scenario. We adopted a pre-processing pipeline with low computational complexity, which can be easily and practically implemented in real-time. Moreover, regression-based continuous fatigue estimation was achieved using power spectral features in conjunction with time as the fatigue label. We sought to compare three regression models and three time windows to demonstrate their effects on the performance of fatigue estimation. Dynamic time warping was proposed as a new measure for evaluating the performance of fatigue estimation. The results derived from the validation of the proposed framework on 19 subjects showed that our proposed framework was promising towards practical implementation. Fatigue estimation by the support vector regression with radial basis function kernel and 5-second window length achieved the best performance. We also provided a comprehensive analysis on the spatial distribution of channels and frequency bands mostly contributing to fatigue estimation, which can inform the feature and channel reduction for real-time fatigue monitoring in practical driving. After reducing the number of electrodes by 75%, the proposed framework retained comparable performance in fatigue estimation. This study demonstrates the feasibility and adaptability of our proposed framework in practical implementation of mental fatigue estimation.

show abstract

“…A study using directed transfer function (DTF) revealed impaired parietalto-frontal coupling in alpha band and enhanced frontal-tocenter coupling in beta band in the left hemisphere [17]. Characteristic path length (L) and the normalized L computed from partial directed coherence in lower alpha band (8)(9)(10) increased significantly during fatigue, indicating an increasing inefficiency of information processing [21]. Such inefficiency during drowsiness was also observed in the delta and theta bands [20].…”

Section: Introductionmentioning

confidence: 99%

Between-Frequency Topographical and Dynamic High-Order Functional Connectivity for Driving Drowsiness Assessment

Harvy

Thakor

Bezerianos

et al. 2019

IEEE Trans. Neural Syst. Rehabil. Eng.

Self Cite

View full text Add to dashboard Cite

Previous studies exploring driving drowsiness utilized spectral power and functional connectivity without considering between-frequency and more complex synchronizations. To complement such lacks, we explored interregional synchronizations based on the topographical and dynamic properties between frequency bands using high-order functional connectivity (HOFC) and envelope correlation. We proposed the dynamic interactions of HOFC, associated-HOFC, and a global metric measuring the aggregated effect of the functional connectivity. The EEG dataset was collected from 30 healthy subjects, undergoing two driving sessions. The two-session setting was employed for evaluating the metric reliability across sessions. Based on the results, we observed reliably significant metric changes, mainly involving the alpha band. In HOFC θα , HOFC αβ , associated-HOFC θα , and associated-HOFC αβ , the connection-level metrics in frontal-central, central-central, and central-parietal/occipitalareas were significantly increased, indicating a dominance in the central region. Similar results were also obtained in the HOFC θαβ and aHOFC θαβ. For dynamic-low-order-FC and dynamic-HOFC, the global metrics revealed a reliably significant increment in the alpha, theta-alpha, and alpha-beta bands. Modularity indexes of associated-HOFC α and associated-HOFC θα also exhibited reliably significant differences. This paper demonstrated that within-band and betweenfrequency topographical and dynamic FC can provide complementary information to the traditional individual-band LOFC for assessing driving drowsiness.

show abstract

Performance Improvement of Driving Fatigue Identification Based on Power Spectra and Connectivity Using Feature Level and Decision Level Fusions

Cited by 12 publications

References 17 publications

EEG-Based Multiclass Workload Identification Using Feature Fusion and Selection

EEG-Based Multiclass Workload Identification Using Feature Fusion and Selection

Regression-Based Continuous Driving Fatigue Estimation: Toward Practical Implementation

Between-Frequency Topographical and Dynamic High-Order Functional Connectivity for Driving Drowsiness Assessment

Contact Info

Product

Resources

About