Guess What?: Towards Understanding Autism from Structured Video Using Facial Affect.

Area of Science:

• Pediatric neurology
• Computational psychology
• Human-computer interaction

Background:

• Autism Spectrum Disorder (ASD) affects 1 in 59 children in the US.
• Current diagnostic and care delivery methods for ASD face challenges, necessitating innovative approaches.
• Mobile tools and wearable devices offer potential for out-of-clinic interventions and research.

Purpose of the Study:

• To present a mobile game, "Guess What?", for acquiring structured video data from children with ASD.
• To apply face tracking and emotion recognition to analyze behavioral data.
• To demonstrate the quantification of engagement and facial affect for future therapeutic applications.

Main Methods:

• Development of a mobile charades-style game, "Guess What?", for data acquisition.
• Utilizing face tracking and emotion recognition algorithms on acquired video data.
• Analyzing facial affect in response to game prompts to measure engagement.

Main Results:

• The "Guess What?" game successfully acquired structured, emotive video data from children with ASD.
• Real-time image processing algorithms effectively quantified engagement and facial affect.
• Specific game prompts and categories were identified as most effective for eliciting diverse emotions.

Conclusions:

• The "Guess What?" system demonstrates efficacy in deriving emotive video data from children with ASD via a gamified mobile platform.
• This approach represents a significant first step towards at-home screenings and outcome measures for ASD.
• The findings support the potential of mobile technology for advancing behavioral disease research and therapies for ASD.