Related Experiment Video
Updated: Jun 13, 2025
07:16
Extraction of the EPP Component from the Surface EMG
Published on: December 16, 2009
12.5K
Real-time identification of noise type contaminated in surface electromyogram signals using efficient statistical
Pornchai Phukpattaranont1, Nantarika Thiamchoo1, Paramin Neranon2
1Department of Electrical and Biomedical Engineering, Faculty of Engineering, Prince of Songkla University, 90110, Songkhla, Thailand.
Medical Engineering & Physics
|September 16, 2024
Summary
This study introduces an efficient real-time system to identify clean and noisy surface electromyogram (EMG) signals. The system accurately detects electrocardiogram, spike, and power line noise using statistical features.
Area of Science:
- Biomedical Engineering
- Signal Processing
- Machine Learning
Background:
- Surface electromyogram (EMG) signal quality is crucial for accurate interpretation.
- Noise contamination, such as electrocardiogram (ECG) interference, spike noise, and power line interference, significantly degrades EMG signal recognition.
- Effective noise removal requires accurate identification of the noise type present in the EMG signal.
Purpose of the Study:
- To develop and validate a real-time, efficient system for classifying clean EMG signals versus those contaminated by specific noise types.
- To investigate the efficacy of statistical descriptors, kurtosis and skewness, for noise identification in EMG signals.
- To propose a computationally efficient method for calculating kurtosis and skewness to reduce processing time and memory requirements.
Main Methods:
- Utilized kurtosis and skewness as input features for a cascading Quadratic Discriminant Analysis (QDA) classifier.
- Developed an optimized algorithm for calculating kurtosis and skewness, reducing computational load.
- Implemented and tested the system in real-time using an ATmega 2560 microcontroller.
Main Results:
- The proposed efficient calculation methods for kurtosis and skewness exhibited low root mean square errors (0.08 and 0.09, respectively) compared to traditional techniques.
- The QDA classifier achieved a high identification accuracy of 96.00% with five-fold cross-validation.
- The system demonstrated real-time capability in distinguishing clean EMG signals from those contaminated with ECG, spike, or power line noise.
Conclusions:
- The proposed real-time system effectively identifies different types of noise in EMG signals.
- The simplified kurtosis and skewness calculations offer a computationally efficient approach for EMG noise analysis.
- This method provides a robust foundation for advanced EMG signal processing and interpretation in clinical and research settings.
