Smile
Bilateral raising of the lip corners. Distinguishes variants with, without, or with undetermined participation of AU6, without classifying the smile as genuine or fake.
FROID Psique / Facial maps
FROID groups combinations of movements into seven families of compatible patterns. The names help organize the reading; they do not conclude what the person feels.
A single word does not tell a whole story. In the same way, a facial movement needs to be understood within the conversation and the professional assessment.
Action Unit estimates describe components of movement. Confirming the same variant across distinct frames records recurrence — not emotional certainty.
Bilateral raising of the lip corners. Distinguishes variants with, without, or with undetermined participation of AU6, without classifying the smile as genuine or fake.
Lowering of the lip corners associated with eyebrow or cheek movements, depending on the combinations available.
Eyebrow and eyelid movements with lip pressure. Variants that depend on unavailable measurements are not presumed.
Eyelid opening associated with horizontal stretching of the lips, which may include other movements of the configuration.
Raised eyebrows, eyelid opening and jaw opening. If another family also matches, the ambiguity is preserved.
Nasal or upper-lip movements associated with the lower-movement proxy used by the engine.
Strictly unilateral pulling or dimple formation. The observed side is recorded; asymmetry does not prove contempt.
The same configuration can occur in different contexts. Not every variant has enough measurements to be determined.
Coefficients received, laterality and the availability of measurements are preserved. Expired readings do not represent the present; repeated frames do not confirm persistence.
A reference image helps place the vocabulary. The thumbnail enlarges on click.
These are FACS reference charts, not FROID's output. What the engine calculates is in the sections above: 16 Action Unit fields, 15 with approximate mapping and AU23 not determined, organized into seven families. FROID does not assess facial-vocal dissonance. Click the image to enlarge.
There is no conclusion about lying, authenticity, intent or conflict between voice and face. Nominal capture at close to three frames per second does not offer certified microexpression detection. Pose, lighting, occlusion and speech affect the reading.