Determination of Emotional State through Physiological Measurement
MetadataShow full metadata
The goal of this thesis is to develop and evaluate methods of emotional response classification using human physiological data. With the continued development of automated systems that interact closely with humans in a more natural manner the ability of such systems to determine the emotional state of nearby subjects and adapt accordingly is increasingly important. Applications include the broad area of affective computing as well as more specific areas such as evaluating the effectiveness of virtual reality based treatment for social phobias. In this work, various non-invasive sensors are used to collect physiological data during virtual reality simulations. Feature extraction, feature selection, and machine learning is performed on the data to determine which signals and algorithms produce the most accurate classification of the subject’s emotional response to the simulations.