A new training approach for deep learning in EEG biometrics using triplet loss and EMG-driven additive data augmentation

Seha, Sherif Nagib Abbas; Hatzinakos, Dimitrios

doi:10.1016/j.neucom.2022.02.084

Cited by 5 publications

(2 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To address the challenge of data scarcity, researchers have focused on data augmentation as a potential solution, aiming to generate augmented data (satisfying the vicinal distribution) via the modification or recombination of existing data (satisfying the empirical distribution) according to specific inductive bias [21]- [23]. In the field of EEG, the data augmentation methods can be classified based on the information used for modification or recombination, including temporal [3], [5], [24]- [26], spatial [3], [27], spectral [26], [28], and component information [29], [30]. However, the existing work on data augmentation with EEG data is still limited.…”

Section: Introductionmentioning

confidence: 99%

Spatial Variation Generation Algorithm for Motor Imagery Data Augmentation: Increasing the Density of Sample Vicinity

Qin,

Yang,

Huang

et al. 2023

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

The imbalanced development between deep learning-based model design and motor imagery (MI) data acquisition raises concerns about the potential overfitting issuemodels can identify training data well but fail to generalize test data. In this study, a Spatial Variation Generation (SVG) algorithm for MI data augmentation is proposed to alleviate the overfitting issue. In essence, SVG generates MI data using variations of electrode placement and brain spatial pattern, ultimately elevating the density of the raw sample vicinity. The proposed SVG prevents models from memorizing the training data by replacing the raw samples with the proper vicinal distribution. Moreover, SVG generates a uniform distribution and stabilizes the training process of models. In comparison studies involving five deep learning-based models across eight datasets, the proposed SVG algorithm exhibited a notable improvement of 0.021 in the area under the receiver operating characteristic curve (AUC). The improvement achieved by SVG outperforms other data augmentation algorithms. Further results from the ablation study verify the effectiveness of each component of SVG. Finally, the studies in the control group with varying numbers of samples show that the SVG algorithm consistently improves the AUC, with improvements ranging from approximately 0.02 to 0.15.

show abstract

Section: Introductionmentioning

confidence: 99%

Spatial Variation Generation Algorithm for Motor Imagery Data Augmentation: Increasing the Density of Sample Vicinity

Qin,

Yang,

Huang

et al. 2023

IEEE Trans. Neural Syst. Rehabil. Eng.

View full text Add to dashboard Cite

show abstract

“…Lawhern et al ( 2018 ) proposed a specific CNN, EEGNet, for the purpose of processing EEG signals. Several EEG-based authentication studies (Kumar et al, 2021 ; Seha and Hatzinakos, 2022 ) have applied EEGNet.…”

Section: Introductionmentioning

confidence: 99%

Authentication using c-VEP evoked in a mild-burdened cognitive task

Huang,

Liao,

et al. 2023

Front. Hum. Neurosci.

View full text Add to dashboard Cite

In recent years, more and more researchers are devoting themselves to the studies about authentication based on biomarkers. Among a wide variety of biomarkers, code-modulated visual evoked potential (c-VEP) has attracted increasing attention due to its significant role in the field of brain-computer interface. In this study, we designed a mild-burdened cognitive task (MBCT), which can check whether participants focus their attention on the visual stimuli that evoke c-VEP. Furthermore, we investigated the authentication based on the c-VEP evoked in the cognitive task by introducing a deep learning method. Seventeen participants were recruited to take part in the MBCT experiments including two sessions, which were carried out on two different days. The c-VEP signals from the first session were extracted to train the authentication deep models. The c-VEP data of the second session were used to verify the models. It achieved a desirable performance, with the average accuracy and F1 score, respectively, of 0.92 and 0.89. These results show that c-VEP carries individual discriminative characteristics and it is feasible to develop a practical authentication system based on c-VEP.

show abstract

Affective EEG-based cross-session person identification using hierarchical graph embedding

Liu,

Jin,

Liu

et al. 2024

Cogn Neurodyn

View full text Add to dashboard Cite

A new training approach for deep learning in EEG biometrics using triplet loss and EMG-driven additive data augmentation

Cited by 5 publications

References 34 publications

Spatial Variation Generation Algorithm for Motor Imagery Data Augmentation: Increasing the Density of Sample Vicinity

Spatial Variation Generation Algorithm for Motor Imagery Data Augmentation: Increasing the Density of Sample Vicinity

Authentication using c-VEP evoked in a mild-burdened cognitive task

Affective EEG-based cross-session person identification using hierarchical graph embedding

Contact Info

Product

Resources

About