The project Stylized Animations for Research on Autism (SARA) aims to develop a better understanding of the cognitive processes behind emotional categorization in children and adolescents with high-functioning autism spectrum disorder (ASD), in comparison with neurotypically developed peers. To this end, we combine novel real-time non-photorealistic rendering (NPR) algorithms, emotional facial animations, and eye-tracking technologies in a framework that serves as an interactive testbed for empirical research. In this paper, we focus on three experiments that: (1) validate real-time facial animations of virtual characters, (2) evaluate the NPR algorithms to create abstracted facial expressions, and (3) elucidate the relation between eye gaze behavior, ASD and alexithymia (i.e., difficulties in expressing ones emotions). The results show that our animations can be used in the proposed experiments; however, more evaluation is needed regarding the NPR abstractions, especially with individuals with ASD. Finally, even though no correlation was found between gaze behavior, ASD and alexithymia, the study opened several questions that will be addressed in future work.

To investigate the difficulties in communication, socialization abilities, and emotion perception deficits in individuals with autism spectrum disorders (ASD), we propose the project SARA. Its main goal is to assess how abstracted emotional facial expressions influence the categorization of the emotions by children and adolescents with high-functioning ASD. This paper focuses on the first pilot study where an adapted version of the Dynamic Emotional Categorization Test (DECT) was implemented. The results support the validity of the emotional facial animations, which is confirmed by the low performance of subjects with high-functioning ASD when categorizing emotions, in comparison with their peers without ASD.

Before the end of the XX century, the inhabitants of this little Planet Earth could meet each other, in general, only after a big effort.