Efficient EOG-based movement classification in IoMT using machine learning algorithms for people with motor...

Area of Science:

• Biomedical Engineering
• Assistive Technology
• Signal Processing

Background:

• Motor disabilities significantly impact independence and quality of life.
• Smart home technologies offer potential for enhanced autonomy.
• Effective human-computer interfaces are crucial for assistive systems.

Purpose of the Study:

• To develop and evaluate an Internet of Medical Things (IoMT) platform for real-time smart home control using electrooculography (EOG).
• To enable patients with motor disabilities to interact with and manage their home environment.
• To investigate the efficacy of signal processing and machine learning techniques for EOG-based control.

Main Methods:

• Utilized electrooculography (EOG) signals for user interaction.
• Applied Stockwell transform (S-transform) and wavelet transform for EOG signal analysis.
• Employed Daubechies (db4) and Symlets (Sym4) wavelet families for feature extraction.
• Classified eye movements using Support Vector Machines (SVM), Kernel Neural Networks (KNN), Ensemble Tree (ET), and Convolutional Neural Networks (CNN).

Main Results:

• Achieved a high average accuracy of 97.7% with the SVM classifier and db4 wavelet.
• Demonstrated superior performance compared to previous methods.
• The db4 wavelet provided better results than the Sym4 wavelet for EOG signal classification.

Conclusions:

• The proposed IoMT platform effectively utilizes EOG signals for smart home control in patients with motor disabilities.
• The combination of S-transform, wavelet transform (specifically db4), and SVM offers a robust and accurate solution.
• This technology has significant potential to improve the independence and quality of life for individuals with motor impairments.