FACS is a taxonomy developed by Paul Ekman and Wallace Friesen that decomposes facial movement into discrete Action Units, each corresponding to a specific muscle or muscle group. It describes what a face is doing without asserting what the person feels.
What it is
FACS catalogues around 44 Action Units — AU12 is the lip-corner puller, AU4 the brow lowerer. Trained human coders can annotate video frame by frame; modern systems detect Action Units automatically from landmark positions.
The important distinction is that FACS itself is descriptive. It says a brow lowered. The step from there to 'this person is angry' is a separate inferential model, and it is that second step which carries the scientific controversy and the regulatory weight.
Why it matters
FACS is the foundation of essentially every commercial facial expression analysis system, which means its assumptions propagate everywhere. Understanding that the coding layer and the emotion-inference layer are different things is the key to reading vendor claims accurately.
A common confusion
FACS is often described as an emotion detection system. It is not — it is a movement taxonomy. Systems built on it add an emotion classifier, and that classifier is where accuracy claims should be scrutinised.
Related
See also emotion inference, emotion ai and baseline calibration. See the full glossary for the rest, or the limits of facial emotion recognition for the wider context.
See where your content loses people
Book a walkthrough and we will show you the engagement data on your own content.