Behavioral · FaceLive

Facial keypoints, expression, and affect from video

ConductVision produces 68-point facial landmarks and the action-unit and expression cues built on them — auditable overlays and frame-level data for attention, affect, and human-factors protocols.

conductvisionFacial expression & affect
68-point facial landmarks
What it measures

Objective, exportable metrics

68 facial landmarks

jaw, brows, nose, eyes, lips

Action-unit cues

facial-movement features over time

Expression dynamics

onset, peak, and offset timing

Blink & gaze

blink rate and gaze direction

Digitizes a validated assessment

Facial-action observational coding

Replaces manual facial-action coding by raters. Output the score the field already trusts, plus the richer digital metrics behind it — and the recording stays auditable.

Workflow

From video to exportable metrics

1

Record the face

Film a clear frontal view of the face during the protocol.

2

Track landmarks

MMPose wholebody facial landmarks are estimated per frame with overlays.

3

Export expression data

Landmarks, action-unit cues, and expression timing export as JSON for analysis.

Research applications

Where teams use it

Attention and expression protocols
Affective-science research
Human-factors and usability studies
Clinical-research facial-symmetry tracking

Facial expression & affect — methods & FAQ

Bring facial expression & affect into your ConductVision workflow

Send a representative video or schedule a consultation to map the outputs your protocol needs.