Electrooculogram filtering using wavelet and wavelet packet transforms.

Area of Science:

• Biomedical Engineering
• Signal Processing
• Neuroscience

Background:

• Electrooculogram (EOG) signals are crucial for studying eye movements and related neurological conditions.
• Traditional digital filters often struggle to effectively remove noise from EOG signals without distorting valuable data.
• Optimizing EOG signal filtering is essential for accurate diagnostic and research applications.

Purpose of the Study:

• To propose and evaluate novel filtering techniques for electrooculogram (EOG) signals.
• To compare the efficacy of Discrete Wavelet Transform (DWT) and Discrete Wavelet Packet Transform (DWPT) against conventional digital filters.
• To identify optimal parameters for wavelet-based EOG signal processing.

Main Methods:

• Implementation of Discrete Wavelet Transform (DWT) for EOG signal filtering.
• Application of Discrete Wavelet Packet Transform (DWPT) for EOG signal denoising.
• Comparative analysis using metrics: signal-to-noise ratio (SNR) improvement, energy preservation, correlation coefficient, and mean absolute error (MAE) reduction.
• Evaluation against standard digital filtering techniques.

Main Results:

• Wavelet and wavelet packet transforms demonstrated substantial improvements in signal-to-noise ratio compared to digital filters.
• The proposed wavelet-based methods achieved significant reduction in mean absolute error.
• Energy preservation and correlation coefficients were favorably impacted by the wavelet approaches.
• Optimal parameters for DWT and DWPT were identified for effective EOG filtering.

Conclusions:

• Discrete Wavelet Transform and Discrete Wavelet Packet Transform offer superior performance for filtering electrooculogram signals.
• Wavelet-based filtering significantly enhances EOG signal quality, outperforming traditional digital filters.
• The findings provide practical guidelines for selecting appropriate wavelet parameters in EOG analysis.