Emotional eye movement markup language for virtual agents

EEMML (Emotional Eye Movement Markup Language) is a scripting tool that enables authors to describe and generate emotional eye movement in virtual agents. The EEMML is capable of describing and generating both basic eye movement and emotional eye movement, including primary (joy, sadness, anger, fear, disgust and surprise) and intermediate (emotions that can be represented as the mixture of two primary emotions) emotions for virtual agents. The emotional eye movement generation framework is based upon the MPEG-4 FAP (facial animation parameters), and the animations are driven by parameters picked from the CohnKanade AU-Coded facial expression database as well as realtime eye movement data(pupil size, blink rate and saccade).