Emotional state detection based on EMG and EOG biosignals: A short survey

With its acquisition of muscular potentials linked to displays of affective state, facial Electromyography (EMG) is the most fitting physiological signal for performing emotional expressions detection. EMG and computer vision-based facial recognition have very different strengths and weaknesses, but their information may complement each other. Additionally, Electrooculographic (EOG) signals can enable gaze tracking and detection of eye movements, but blinks and saccades cannot directly indicate emotional state. In this paper, we survey and analyze the methods, strengths and challenges of using these biosignal sources for detecting emotional states, and then we propose a framework that combines simultaneously facial expression detection, using EMG, with saccade detection using EOGs, to classify four basic expressions: neutral, sad, happy, and angry.