Daily Arxiv

This page organizes papers related to artificial intelligence published around the world.
This page is summarized using Google Gemini and is operated on a non-profit basis.
The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.

Real-time Noise Detection and Classification in Single-Channel EEG: A Lightweight Machine Learning Approach for EMG, White Noise, and EOG Artifacts

Created by
  • Haebom

Author

Hossein Enshaei, Pariya Jebreili, Sayed Mahmoud Sakhaei

Electroencephalogram (EEG) Artifact Detection: A Hybrid Spectral-Temporal Framework

Outline

This paper proposes a hybrid spectral-temporal framework for real-time detection and classification of eye (EOG), muscle (EMG), and white noise artifacts from single-channel EEG. This framework combines time-domain low-pass filtering and frequency-domain power spectral density (PSD) analysis to extract features and minimizes redundancy through principal component analysis (PCA)-based feature fusion. Leveraging a lightweight multilayer perceptron (MLP) architecture, it achieves higher accuracy than advanced convolutional neural networks (CNNs) and recurrent neural networks (RNNs) and outperforms simultaneous multi-source contamination (EMG, EOG, and white noise). Its short training time and robust performance across a wide range of signal-to-noise ratio (SNR) environments enhance the potential for real-time applications in wearable brain-computer interfaces (BCIs).

Takeaways, Limitations

•
Takeaways:
◦
Achieving both accuracy and computational efficiency in single-channel EEG artifact detection.
◦
We demonstrate that domain knowledge-based feature fusion can outperform complex deep learning models.
◦
Addressing the problem of simultaneous multi-source contamination, improving applicability in real-world environments.
◦
Delivering fast, robust performance suitable for real-time BCI applications.
•
Limitations:
◦
Limited to single-channel EEG data, lack of generalization to multi-channel EEG data.
◦
Further validation is needed to ensure that the performance of the proposed framework remains consistent across different artifact types or complex noise environments.
◦
Possibility of information loss due to dimensionality reduction when PCA-based feature fusion.
👍
Made with Slashpage