What kinds of motion can Kalidoface 3D track?

Updated

Kalidoface 3D tracks three categories of motion through a browser webcam: facial expressions, full-body pose, and finger movement. These are combined in real time to drive a 3D character, so the character mirrors your face, body, and hands without any dedicated hardware beyond a camera.

Face tracking

Face tracking captures the movements and expressions of your head and face. In practice this covers:

  • Head orientation — turning, tilting, and nodding
  • Eyebrow and eye movement, including blinks
  • Mouth shapes such as opening, closing, and smiling

This is what makes the character's face react as you talk or emote, rather than staying frozen while only the body moves.

Full body pose tracking

Pose tracking follows the position and orientation of your body's major joints so the character can copy your stance and gestures. This typically includes:

  • Torso and shoulder position
  • Arm and elbow movement
  • Leg and knee movement

Because it reads your overall posture, you can make the character sit, lean, wave, or shift weight simply by doing it yourself in front of the camera.

Finger tracking

Finger tracking adds detail at the hand level, capturing individual finger positions so the character can form gestures and hand poses. This is the finest-grained of the three and is what separates a simple body mirror from a character that can point, grip, or make specific hand signs.

How the three combine

The three tracking layers run together and feed a single 3D character:

Tracking type What it reads What it drives
Face Head angle, eyes, brows, mouth Character's facial expression and head movement
Full body pose Torso, arms, legs, joints Character's stance and body gestures
Fingers Individual finger positions Character's hand poses and gestures

The result is one animated character whose face, body, and hands all follow you at once. You do not switch between modes — the layers stack so the character stays consistent across expression, posture, and gesture.

What you need for it to work

All three tracking types rely on the browser webcam, so the practical requirements are the same for each:

  • A working webcam the browser can access
  • Permission granted for the site to use the camera
  • Enough light and a clear view of your face, body, and hands for the tracker to read them

If one layer seems unresponsive, the cause is usually the camera view rather than the tracking type itself — for example, hands out of frame will not register finger movement even when face and body tracking are working.