FACS Action Unit Reference Grid AI Image Prompt

Type: FACS facial action coding system reference grid poster Title: FACS Reference Grid — 48 Action Units (v3 Step 2/3 fix) Subtitle: OLD algorithm woman, Cafe Window | Model Name: banana 2 | eyes v3.1, flash mug-preview Overall style: dark technical contact-sheet layout, small white sans-serif typography, thin grid gutters, photographic thumbnails, warm cafe-window lighting, muted black and charcoal UI background Subject: young woman with long dark hair holding a beige ceramic cup near her mouth Setting: warm cafe window interior with vertical window frames and golden daylight Face treatment: each thumbnail shows the same portrait composition with a large soft-edged beige rectangular anonymization mask covering the central face area while hair, hands, cup, neck, and background remain visible Layout: canvas is portrait poster, approximately 4:5 aspect ratio, black background; header is two-line header at top left, compact and technical; grid has 49 total tiles, 48 filled action units, 1 empty tile, 7 columns, 7 rows; thumbnail style: each tile has a photographic image on top and a dark caption bar below with a numeric index, AU code, and action label; empty tile: bottom-right tile is blank dark gray with tiny text reading missing and a label line 49. --- Action units: (48 items, each with index, code, and label, omitted for brevity) Generation instructions: Create a consistent reference grid using the same subject, camera angle, cup, lighting, and background in every filled tile. Vary only the facial action or head/eye direction indicated by each label, even though the face is mostly hidden by the anonymization mask. Keep labels crisp and legible, with exact AU codes and numbering.