Automatic Recognition of Posed Facial Expression of Emotion in Individuals with Autism Spectrum Disorder.

Rabin et al. (2019) asked people with and without autism to pose happy, sad, angry, and surprised faces. A computer program called FACET watched each video and counted tiny facial muscle movements.

The same parents filled out a social-skills checklist about their child. The team wanted to know if the computer scores matched real-life social ability.

The autistic group moved different face muscles than the typical group. Their posed smiles, frowns, and scowls looked slightly off even though they were trying.

Kids whose faces differed the most also had lower parent-rated social skills. The computer numbers tracked everyday social problems.

Lemons et al. (2015) used human judges and saw the same thing: autistic adults posed stranger but still readable faces. Joseph adds FACET software, so you can now measure change without paying raters.

Kuang et al. (2025) pooled 25 brain-scan papers and found weaker left inferior frontal gyrus activity during emotion tasks. Joseph shows the output side—different muscle activation—linking brain and face data.

Libero et al. (2016) tracked kids for 18 weeks and proved facial emotion recognition can improve. Joseph gives you a cheap webcam tool to check if your intervention is actually moving the face muscles.

You can run FACET on a tablet with any webcam. Film your client posing five emotions at intake, then every month. If the action-unit scores move closer to typical patterns, your social-skills program is working. No extra staff, no cost, just 30 seconds of video and free software.