Automatic face analysis tools for interactive digital games

Individuals with Autism Spectrum Condition (ASC) face a number of challenges during social interaction due to their limited ability to decode non-verbal cues. Interactive digital games can help as tools to teach such people how to express and recognise non-verbal cues and interpret their emotional meanings. In this paper, we present a set of interactive interfaces that can assist in teaching people with social communication difficulties about facial expressions and face touches. We describe the underlying technologies of three systems: Facial Affect Mapping Engine (FAME), age and gender estimation module and hand-over-face gesture interactive mimicking avatar. The three systems are fully automatic and run in real-time. We show their potential use as integral components in serious games that focus on social inclusion.