Affect recognition for handheld devices

• Main Authors: Go, Giorgio Ferrero O., Ling, Giselle Odelia C., Uy, Timothy Christian T.
• Format: text
• Language: English
• Published: Animo Repository 2013
• Online Access: https://animorepository.dlsu.edu.ph/etd_honors/372
• Institution: De La Salle University

This study focuses on the development of an affect recognition system for Android handheld devices following the client-server framework, where data acquisition and display of outputs are done on the handheld device while feature extraction and classification are dine on the Java application server running on a computer. Data acquisition is done through a math game where the subject is recorded while playing the game. Motion history images (MHI) and edge orientation histogram (EOH) were used to represent the face. Categorical labels " Interest, Amusement and Neutral " are the emotions used and these are classified through a decision tree model following the C4.5 or J48 algorithm with an accuracy score of 89% and 94%.