26 research outputs found
Optimal accuracy performance in music-based EEG signal using Matthew correlation coefficient advanced (MCCA)
Get PDFThe connection between music and human are very synonyms because music could reduce stress. The state of stress could be measured using EEG signal, an electroencephalogram (EEG) measurement which contains an arousal and valence index value. In previous studies, it is found that the Matthew Correlation Coefficient (MCC) performance accuracy is of 85±5%. The arousal indicates strong emotion, and valence indicates positive and negative degree of emotion. Arousal and valence values could be used to measure the accuracy performance. This research focuses on the enhance MCC parameter equation based on arousal and valence values to perform the maximum accuracy percentage in the frequency domain and time-frequency domain analysis. Twenty-one features were used to improve the significance of feature extraction results and the investigated arousal and valence value. The substantial feature extraction involved alpha, beta, delta and theta frequency bands in measuring the arousal and valence index formula. Based on the results, the arousal and valance index is accepted to be applied as parameters in the MCC equations. However, in certain cases, the improvement of the MCC parameter is required to achieve a high accuracy percentage and this research proposed Matthew correlation coefficient advanced (MCCA) in order to improve the performance result by using a six sigma method. In conclusion, the MCCA equation is established to enhance the existing MCC parameter to improve the accuracy percentage up to 99.9% for the arousal and valence index
The multimodal parameter enhancement of electroencephalogram signal for music application
Get PDFBlinding of modality has been influenced decision of multimodal in several circumstances. Sometimes, certain electroencephalogram (EEG) signal is omitted to achieve the highest accuracy of performance. Therefore, the aim for this paper is to enhance the multimodal parameters of EEG signals based on music applications. The structure of multimodal is evaluated with performance measure to ensure the implementation of parameter value is valid to apply in the multimodal equation. The modalities’ parameters proposed in this multimodal are weighted stress condition, signal features extraction, and music class. The weighted stress condition was obtained from stress classes. The EEG signal produces signal features extracted from the frequency domain and time-frequency domain via techniques such as power spectrum density (PSD), short-time Fourier transform (STFT), and continuous wavelet transform (CWT). Power value is evaluated in PSD. The energy distribution is derived from STFT and CWT techniques. Two types of music were used in this experiment. The multimodal fusion is tested using a six-performance measurement method. The purposed multimodal parameter shows the highest accuracy is 97.68%. The sensitivity of this study presents over 95% and the high value for specificity is 89.5%. The area under the curve (AUC) value is 1 and the F1 score is 0.986. The informedness values range from 0.793 to 0.812 found in this paper
A ROBUST FRAMEWORK FOR DRIVER FATIGUE DETECTION FROM EEG SIGNALS USING ENHANCEMENT OF MODIFIED Z-SCORE AND MULTIPLE MACHINE LEARNING ARCHITECTURES
Get PDFPhysiological signals, such as electroencephalogram (EEG), are used to observe a driver’s brain activities. A portable EEG system provides several advantages, including ease of operation, cost-effectiveness, portability, and few physical restrictions. However, it can be challenging to analyse EEG signals as they often contain various artefacts, including muscle activities, eye blinking, and unwanted noises. This study utilised an independent component analysis (ICA) approach to eliminate such unwanted signals from the unprocessed EEG data of 12 young, physically fit male participants between the ages of 19 and 24 who took part in a driving simulation. Furthermore, driver fatigue state detection was carried out using multichannel EEG signals obtained from O1, O2, Fp1, Fp2, P3, P4, F3, and F4. An enhanced modified z-score was utilised with features extracted from a time-frequency domain continuous wavelet transform (CWT) to elevate the reliability of driver fatigue classification. The proposed methodology offers several advantages. First, multichannel EEG analysis improves the accuracy of sleep stage detection, which is vital for accurate driver fatigue detection. Second, an enhanced modified z-score in feature extraction is more robust than conventional z-score techniques, making it more effective for removing outlier values and improving classification accuracy. Third, the proposed approach for detecting driver fatigue employs multiple machine learning classifiers, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Artificial Neural Networks (ANNs) that utilise Long Short-Term Memory (LSTM), and also machine learning techniques like Support Vector Machines (SVM). The evaluation of five classifiers was performed through 5-fold cross-validation. The outcomes indicate that the suggested framework attains exceptional precision in identifying driver fatigue, with an average accuracy rate of 96.07%. Among the classifiers, the ANN classifier achieved the most significant precision of 99.65%, and the SVM classifier ranked second with an accuracy of 97.89%. Based on the results of the receiver operating characteristic (ROC) and area under the curve (AUC) analysis, it was observed that all the classifiers had an outstanding performance, with an average AUC value of 0.95. This study’s contribution lies in presenting a comprehensive and effective framework that can accurately detect driver fatigue from EEG signals.
ABSTRAK: Isyarat fisiologi, seperti elektroencefalogram (EEG), digunakan bagi memerhati aktiviti otak pemandu. Sistem EEG mudah alih menyediakan beberapa kelebihan, termasuk kemudahan operasi, keberkesanan kos, mudah alih dan sedikit sekatan fizikal. Namun, isyarat EEG mungkin sukar dianalisis kerana ia sering mengandungi pelbagai artifak, termasuk aktiviti otot, mata berkedip dan bunyi yang tidak diingini. Kajian ini menggunakan pendekatan analisis komponen bebas (ICA) bagi membuang isyarat tidak diperlukan daripada data EEG yang belum diproses daripada 12 peserta lelaki muda, cergas fizikal berumur 19 hingga 24 tahun yang mengambil bahagian dalam simulasi pemanduan. Tambahan, pengesanan keadaan lesu pemandu telah dijalankan menggunakan isyarat EEG berbilang saluran yang diperoleh dari O1, O2, Fp1, Fp2, P3, P4, F3, dan F4. Penambah baik skor z digunakan dengan ciri diekstrak daripada transformasi wavelet berterusan (CWT) domain frekuensi masa bagi meningkatkan kebolehpercayaan klasifikasi keletihan pemandu. Metodologi yang dicadangkan menawarkan beberapa kelebihan. Pertama, analisis EEG berbilang saluran meningkatkan ketepatan pengesanan peringkat tidur, penting bagi pengesanan keletihan pemandu secara tepat. Kedua, penambah baik skor z dalam pengekstrak ciri adalah lebih teguh daripada teknik skor z konvensional, menjadikannya lebih berkesan bagi membuang unsur luaran dan meningkatkan ketepatan pengelasan. Ketiga, pendekatan yang dicadangkan bagi mengesan keletihan pemandu menggunakan pelbagai pengelas pembelajaran mesin, seperti Rangkaian Neural Konvolusi (CNN), Rangkaian Neural Berulang (RNN), Rangkaian Neural Buatan (ANN) yang menggunakan Memori Jangka Pendek Panjang (LSTM), dan juga teknik pembelajaran mesin seperti Mesin Vektor Sokongan (SVM). Penilaian lima pengelas dilakukan melalui pengesahan silang 5 kali ganda. Dapatan kajian menunjukkan cadangan rangka kerja ini mencapai ketepatan yang luar biasa dalam mengenal pasti keletihan pemandu, dengan kadar ketepatan purata 96.07%. Antara kesemua pengelas, pengelas ANN mencapai ketepatan paling ketara sebanyak 99.65%, dan pengelas SVM menduduki tempat kedua dengan ketepatan 97.89%. Berdasarkan keputusan analisis ciri operasi penerima (ROC) dan kawasan di bawah lengkung (AUC), didapati semua pengelas mempunyai prestasi cemerlang, dengan purata nilai AUC 0.95. Sumbangan kajian ini adalah terletak pada rangka kerja yang komprehensif dan berkesan mengesan keletihan pemandu secara tepat melalui isyarat EEG
The multimodal parameter enhancement of electroencephalogram signal for music application
Get PDFBlinding of modality has been influenced decision of multimodal in several circumstances. Sometimes, certain electroencephalogram (EEG) signal is omitted to achieve the highest accuracy of performance. Therefore, the aim for this paper is to enhance the multimodal parameters of EEG signals based on music applications. The structure of multimodal is evaluated with performance measure to ensure the implementation of parameter value is valid to apply in the multimodal equation. The modalities’ parameters proposed in this multimodal are weighted stress condition, signal features extraction, and music class. The weighted stress condition was obtained from stress classes. The EEG signal produces signal features extracted from the frequency domain and time-frequency domain via techniques such as power spectrum density (PSD), short-time Fourier transform (STFT), and continuous wavelet transform (CWT). Power value is evaluated in PSD. The energy distribution is derived from STFT and CWT techniques. Two types of music were used in this experiment. The multimodal fusion is tested using a six-performance measurement method. The purposed multimodal parameter shows the highest accuracy is 97.68%. The sensitivity of this study presents over 95% and the high value for specificity is 89.5%. The area under the curve (AUC) value is 1 and the F1 score is 0.986. The informedness values range from 0.793 to 0.812 found in this paper
A Hybrid Deep Learning Model for Detecting Driver Fatigue Using Electroencephalogram Signals
Get PDFRoad accidents caused by driver fatigue are a significant public safety concern, and detecting driver fatigue is crucial for preventing such incidents. Existing methods for detecting driver fatigue are limited in their effectiveness, and there is a need for more accurate and reliable methods. This study presents a solution to the problem of accurately identifying driver fatigue using electroencephalogram (EEG) signals. The approach involves the development of a hybrid deep learning model that incorporates both a deep belief network (DBN) and a recurrent neural network (RNN). We trained and evaluated our model on a dataset of EEG signals collected from drivers in normal and fatigued states. The effectiveness of the hybrid model in accurately categorizing driver fatigue was evaluated in comparison to two other classifiers. The results of the study indicate that the hybrid model outperformed the other classifiers in terms of accuracy, sensitivity, specificity, precision, and F1 score, suggesting its superior performance. We observed that the model’s accuracy and loss remained consistent even when the number of epochs was low, indicating that the model effectively learned to classify EEG signals and did not overfit the training data. Further evaluation of the hybrid model with varying numbers of epochs revealed that the optimal number for the model was 50. Additionally, analysis of the loss function during training demonstrated that the model effectively learned to classify EEG signals without overfitting the training data. The proposed hybrid model achieved an overall accuracy of 99.98%, with perfect sensitivity (100%) and high specificity (99.95%), precision (99.95%), recall (100%), and F1 score (99.98%). These results indicate that the proposed hybrid deep learning model outperformed the individual DBN and RNN models in classifying EEG signals. Our study’s results demonstrate the potential of the proposed hybrid model to accurately detect driver fatigue, which could contribute to the development of more effective and reliable methods for preventing road accidents caused by driver fatigue
Enhancing driver fatigue detection accuracy in on-road driving systems using an LSTM-DNN hybrid model with modified Z-Score and morlet wavelet
No full textDriver fatigue is a significant safety concern in transportation systems, with the potential to cause accidents. Detecting and addressing driver fatigue in real time is crucial for improving road safety. This research paper introduces an innovative method for detecting driver fatigue using electroencephalogram (EEG) signals, enhanced by the Morlet mother wavelet and modified z-score feature. The Morlet wavelet is adapted to capture both temporal and frequency information from EEG signals associated with driver fatigue, while the modified z-score feature measures abnormal EEG activity. Three deep learning models, Long Short-Term Memory (LSTM), Deep Neural Network (DNN), and LSTM-DNN, are employed to classify the data. The LSTM model captures long-term dependencies, the DNN model learns complex relationships, and the hybrid LSTM-DNN model combines their strengths to improve classification accuracy. The proposed approach demonstrates its effectiveness through comprehensive experiments, achieving high accuracy, specificity, sensitivity, F1-score, and recall in driver fatigue detection. The LSTM-DNN hybrid model showed exceptional performance, achieving an accuracy of 99.99% in classifying EEG signals. This showcases its remarkable precision in accurately categorizing the signals. Additionally, the LSTM-DNN model exhibited a specificity of 99.98% and a sensitivity of 100.00%, indicating its capability to classify driver fatigue states accurately. Furthermore, the F1-score and recall for the LSTM-DNN model were 99.99% and 100.00%, respectively
A Hybrid Deep Learning Model for Detecting Driver Fatigue Using Electroencephalogram Signals
Get PDFRoad accidents caused by driver fatigue are a significant public safety concern, and detecting driver fatigue is crucial for preventing such incidents. Existing methods for detecting driver fatigue are limited in their effectiveness, and there is a need for more accurate and reliable methods. This study presents a solution to the problem of accurately identifying driver fatigue using electroencephalogram (EEG) signals. The approach involves the development of a hybrid deep learning model that incorporates both a deep belief network (DBN) and a recurrent neural network (RNN). We trained and evaluated our model on a dataset of EEG signals collected from drivers in normal and fatigued states. The effectiveness of the hybrid model in accurately categorizing driver fatigue was evaluated in comparison to two other classifiers. The results of the study indicate that the hybrid model outperformed the other classifiers in terms of accuracy, sensitivity, specificity, precision, and F1 score, suggesting its superior performance. We observed that the model’s accuracy and loss remained consistent even when the number of epochs was low, indicating that the model effectively learned to classify EEG signals and did not overfit the training data. Further evaluation of the hybrid model with varying numbers of epochs revealed that the optimal number for the model was 50. Additionally, analysis of the loss function during training demonstrated that the model effectively learned to classify EEG signals without overfitting the training data. The proposed hybrid model achieved an overall accuracy of 99.98%, with perfect sensitivity (100%) and high specificity (99.95%), precision (99.95%), recall (100%), and F1 score (99.98%). These results indicate that the proposed hybrid deep learning model outperformed the individual DBN and RNN models in classifying EEG signals. Our study’s results demonstrate the potential of the proposed hybrid model to accurately detect driver fatigue, which could contribute to the development of more effective and reliable methods for preventing road accidents caused by driver fatigue
A robust framework for driver fatigue detection from EEG signals using enhancement of modified Z-score and multiple machine learning architectures
Get PDFPhysiological signals, such as electroencephalogram (EEG), are used to observe a driver’s brain activities. A portable EEG system provides several advantages, including ease of operation, cost-effectiveness, portability, and few physical restrictions. However, it can be challenging to analyse EEG signals as they often contain various artefacts, including muscle activities, eye blinking, and unwanted noises. This study utilised an independent component analysis (ICA) approach to eliminate such unwanted signals from the unprocessed EEG data of 12 young, physically fit male participants between the ages of 19 and 24 who took part in a driving simulation. Furthermore, driver fatigue state detection was carried out using multichannel EEG signals obtained from O1, O2, Fp1, Fp2, P3, P4, F3, and F4. An enhanced modified z-score was utilised with features extracted from a time-frequency domain continuous wavelet transform (CWT) to elevate the reliability of driver fatigue classification. The proposed methodology offers several advantages. First, multichannel EEG analysis improves the accuracy of sleep stage detection, which is vital for accurate driver fatigue detection. Second, an enhanced modified z-score in feature extraction is more robust than conventional z-score techniques, making it more effective for removing outlier values and improving classification accuracy. Third, the proposed approach for detecting driver fatigue employs multiple machine learning classifiers, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Artificial Neural Networks (ANNs) that utilise Long Short-Term Memory (LSTM), and also machine learning techniques like Support Vector Machines (SVM). The evaluation of five classifiers was performed through 5-fold cross-validation. The outcomes indicate that the suggested framework attains exceptional precision in identifying driver fatigue, with an average accuracy rate of 96.07%. Among the classifiers, the ANN classifier achieved the most significant precision of 99.65%, and the SVM classifier ranked second with an accuracy of 97.89%. Based on the results of the receiver operating characteristic (ROC) and area under the curve (AUC) analysis, it was observed that all the classifiers had an outstanding performance, with an average AUC value of 0.95. This study’s contribution lies in presenting a comprehensive and effective framework that can accurately detect driver fatigue from EEG signals
A hybrid deep learning model for detecting driver fatigue using electroencephalogram signals
Get PDFRoad accidents caused by driver fatigue are a significant public safety concern, and detecting driver fatigue is crucial for preventing such incidents. Existing methods for detecting driver fatigue are limited in their effectiveness, and there is a need for more accurate and reliable methods. This study presents a solution to the problem of accurately identifying driver fatigue using electroencephalogram (EEG) signals. The approach involves the development of a hybrid deep learning model that incorporates both a deep belief network (DBN) and a recurrent neural network (RNN). We trained and evaluated our model on a dataset of EEG signals collected from drivers in normal and fatigued states. The effectiveness of the hybrid model in accurately categorizing driver fatigue was evaluated in comparison to two other classifiers. The results of the study indicate that the hybrid model outperformed the other classifiers in terms of accuracy, sensitivity, specificity, precision, and F1 score, suggesting its superior performance. We observed that the model’s accuracy and loss remained consistent even when the number of epochs was low, indicating that the model effectively learned to classify EEG signals and did not overfit the training data. Further evaluation of the hybrid model with varying numbers of epochs revealed that the optimal number for the model was 50. Additionally, analysis of the loss function during training demonstrated that the model effectively learned to classify EEG signals without overfitting the training data. The proposed hybrid model achieved an overall accuracy of 99.98%, with perfect sensitivity (100%) and high specificity (99.95%), precision (99.95%), recall (100%), and F1 score (99.98%). These results indicate that the proposed hybrid deep learning model outperformed the individual DBN and RNN models in classifying EEG signals. Our study’s results demonstrate the potential of the proposed hybrid model to accurately detect driver fatigue, which could contribute to the development of more effective and reliable methods for preventing road accidents caused by driver fatigue
Enhancement of morlet mother wavelet in time–frequency domain in electroencephalogram (EEG) signals for driver fatigue classification
No full textDriving is hazardous due to various factors, including driving attitudes, road type, and driving perceptual environment. These influences factors may cause a fatigue condition. Moreover, less driving experience and lack of alertness can also be contributed to dangerous accidents. Fatigued driving is a key factor in car accidents worldwide because of sleep disorders and driving durations. An EEG signal is used to determine changes in brain activity for diagnosing driver fatigue states. Artifacts were removed using independent component analysis (ICA) in the preprocessing stage. Then, features are extracted from the temporal region of the brain using eight channels (Fp1, Fp2, O1, O2, F4, F3, P4, and P3). The frequency bands used are alpha, delta, and theta. In continuous wavelet transform analysis, the Morlet wavelet is a fast wavelet transform in time–frequency analysis. Still, it has shift sensitivity and lacks phase information, affecting the frequency resolution analysis. This study proposes the enhancement of the Morlet mother wavelet for frequency resolution in the time–frequency domain using independent component analysis to overcome the drawbacks of the Morlet wavelet. The proposed technique can increase the percentage of driver fatigue classification accuracy of EEG signals. Then, the artificial neural network (ANN) classifier with Levenberg–Marquardt (LM) training algorithm gives the highest accuracy of the classification results with 97.40%, followed by the k-nearest neighbor (KNN) with 95.83% and the support vector machine (SVM) with 83%
