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

In this paper, we present an Internet of Medical Things (IoMT)-based platform that depends on electrooculography (EOG) to assist, control, and monitor a smart home environment in real time for patients with motor disabilities. Users can interact with the intelligent environment through a Graphical User Interface (GUI) that offers predefined options for controlling doors, windows, lights, air conditioning, temperature, and TV functions. The proposed approach is based mainly on the utilisation of two transforms, namely Stockwell transform (S-transform) and wavelet transform, for detection of abrupt changes in EOG signals. Several statistical attributes of the processed EOG signals are utilised to characterise them in order to detect each eye movement. Two different wavelet families, namely Daubechies (db4) and Symlets (Sym4), are considered. Finally, the data is classified using three types of Machine Learning (ML) algorithms in addition to a Deep Learning (DL) algorithm, namely Support Vector Machines (SVM), Kernel Neural Networks (KNN), Ensemble Tree (ET), and Convolutional Neural Networks (CNN) classifiers. The proposed approach reveals the best results in comparison with the results of previous methods. A high average accuracy of 97.7% is achieved with the SVM classifier using the db4 wavelet, while an accuracy of 95.75% is achieved with the Sym4 wavelet, which indicates that the db4 wavelet gives the best results.