An efficient EEGNet processor design for portable EEG-Based BCIs

Feng, Lichen; Yang, Liying; Liu, Shubin; Han, Chenxi; Zhang, Yueqi; Zhu, Zhangming

doi:10.1016/j.mejo.2021.105356

Cited by 17 publications

(3 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ai et al [29] successfully deployed a lightweight convolutional neural network on a TSMC 65nm IP core for EEG-based epilepsy prediction, yielding an impressive accuracy rate of 87.9%. Similarly, Feng et al [30] deployed an EEGNet on a Xilinx KC705 FPGA for four-class classification utilizing event-related potential (ERP) EEG data, resulting in a remarkable classification accuracy of 96.03%.…”

Section: Related Workmentioning

confidence: 99%

An Epileptic EEG Detection Method Based on Data Augmentation and Lightweight Neural Network

Wang,

Liu,

Zhuo

et al. 2024

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

Objective: Epilepsy, an enduring neurological disorder, afflicts approximately 65 million individuals globally, significantly impacting their physical and mental wellbeing. Traditional epilepsy detection methods are labor-intensive, leading to inefficiencies. Although deep learning techniques for brain signal detection have gained traction in recent years, their clinical application advancement is hindered by the significant requirement for high-quality data and computational resources during training. Methods & Results: The neural network training initially involved merging two datasets of different data quality, namely Bonn University datasets and CHB-MIT datasets, to bolster its generalization capabilities. To tackle the issues of dataset size and class imbalance, we employed small window segmentation and Synthetic Minority Over-sampling Technique (SMOTE). algorithms to augment and equalize the data. A streamlined neural network architecture was then proposed, drastically reducing the model's training parameters. Notably, a model trained with a mere 9,371 parameters yielded impressive results. The three-classification task on the combined dataset delivered an accuracy of 98.52%, sensitivity of 97.99%, specificity of 99.35%, and precision of 98.44%.Conclusions: The experimental findings of this study underscore the superiority of the proposed method over existing approaches in both model size reduction and accuracy enhancement. As a result, it is more apt for deployment in low-cost, low computational hardware devices, including wearable technology, and various clinical applications.

show abstract

Section: Related Workmentioning

confidence: 99%

An Epileptic EEG Detection Method Based on Data Augmentation and Lightweight Neural Network

Wang,

Liu,

Zhuo

et al. 2024

IEEE J. Transl. Eng. Health Med.

View full text Add to dashboard Cite

show abstract

“…In this project a recently developed emotional detection structure dependent on a recording feed in real time is done by employing a support vector machine (SVM) which is a reliable classification algorithm [8]. In this project a deep convolutional neural network model, EEGNet, Vivado and its hardware implementation, has been evolved for obtaining generalization towards different BCI Paradigms for design in portable EEG based BCI's [9]. The technique used in this work is a future trail using SVM to distribute emotions and modern approach of preprocessing in the structure of local secular pattern in consequence by PCA were used to feed to SVM classifier model [10].…”

Section: Related Workmentioning

confidence: 99%

Classification of Emotion using Eeg Signals: an FPGA Based Implementation

B D,

B S,

Totiger

et al. 2023

IJRTE

View full text Add to dashboard Cite

An electroencephalograph is a device that records all electrical energy in the human brain using wearable metal electrodes placed on the skull. Electrical impulses connect brain cells and are always mobile, even at rest. This activity appears as a squiggly line in EEG recordings. Activity gaze data is pre-processed to a frequency range of 0 to 75 Hz. This creates a new matrix with a sample rate of 200 Hz and a range of 0-75 Hz. A finite-impulse-response low-pass filter was used because the bandpass would distort his EEG data after processing. Each pre-processed EEG signal has an output, which completes feature extraction. Principal Component Analysis or PCA is passed in the feature reduction phase. PCA is an analytical process that uses singular value decomposition to transform a collection of corresponding features into mutually uncorrelated features or principal components. Principal component analysis: (a) mean normalization of features (b) covariance matrix (c) eigenvectors (d) reduced features or principal components. The above steps are passed to the SVM classifier for sentiment output. His VHDL code and testbench for 2*2 matrices were written, waveforms and RTL schemes were created in Xilinx 14.5. For the FPGA implementation, a Simulink model was designed, and the eigenvalues were pre-determined using a system generator.

show abstract

“…In this sense, Zhu et al [27] developed an ensemble learning coupled to the EEGNet network to improve the ear-EEG signals' classification for SSVEP-based BCI, achieving an accuracy of 81.74%. Lastly, Feng et al [28] implemented a real-time EEGNet classifier on an FPGA board, using only 2.54% of the board's resources and consuming 3.66% of the maximum power available. Similarly, Tsukahara et al [29] achieved an accuracy of 88.75%, implementing the EEGNet architecture on a Virtex-7 FPGA platform to classify EEG data from the MNE dataset.…”

Section: Introductionmentioning

confidence: 99%

Motor Imagery Multi-Tasks Classification for BCIs Using the NVIDIA Jetson TX2 Board and the EEGNet Network

Mwata-Velu

Niyonsaba-Sebigunda

Aviña-Cervantes

et al. 2023

Sensors

View full text Add to dashboard Cite

Nowadays, Brain–Computer Interfaces (BCIs) still captivate large interest because of multiple advantages offered in numerous domains, explicitly assisting people with motor disabilities in communicating with the surrounding environment. However, challenges of portability, instantaneous processing time, and accurate data processing remain for numerous BCI system setups. This work implements an embedded multi-tasks classifier based on motor imagery using the EEGNet network integrated into the NVIDIA Jetson TX2 card. Therefore, two strategies are developed to select the most discriminant channels. The former uses the accuracy based-classifier criterion, while the latter evaluates electrode mutual information to form discriminant channel subsets. Next, the EEGNet network is implemented to classify discriminant channel signals. Additionally, a cyclic learning algorithm is implemented at the software level to accelerate the model learning convergence and fully profit from the NJT2 hardware resources. Finally, motor imagery Electroencephalogram (EEG) signals provided by HaLT’s public benchmark were used, in addition to the k-fold cross-validation method. Average accuracies of 83.7% and 81.3% were achieved by classifying EEG signals per subject and motor imagery task, respectively. Each task was processed with an average latency of 48.7 ms. This framework offers an alternative for online EEG-BCI systems’ requirements, dealing with short processing times and reliable classification accuracy.

show abstract

An efficient EEGNet processor design for portable EEG-Based BCIs

Cited by 17 publications

References 19 publications

An Epileptic EEG Detection Method Based on Data Augmentation and Lightweight Neural Network

An Epileptic EEG Detection Method Based on Data Augmentation and Lightweight Neural Network

Classification of Emotion using Eeg Signals: an FPGA Based Implementation

Motor Imagery Multi-Tasks Classification for BCIs Using the NVIDIA Jetson TX2 Board and the EEGNet Network

Contact Info

Product

Resources

About