NV. FACE LABVISION ENGINE 02 NV SHARP SOLUTIONS
INTERACTIVE COMPUTER VISION / 002

REALITY,
REWIRED.

Your movement becomes a digital universe. Track your face, change the color of a minimal top-left accent with hand gestures, and view a clean face mesh — entirely inside your browser.

01 / FACE MESH02 / MINIMAL ACCENT03 / LIVE ART
DESIGNED & DEVELOPED BYNavod Ravindu Jayaweera
01 / LIVE EXPERIENCE
INTERACTIVE PLAYGROUND
SYSTEM STANDBY
NV / VISION ENGINE 02
LIVE COMPUTER VISION
THE FUTURE IS LOOKING AT YOU.

Enable your camera to turn this abstract sculpture into a responsive face mesh.

X — / Y —
AWAITING CAMERA INPUT
-- TRACK/S
VISION ENGINE / 02Minimal accent. Clear face.
One minimal accent · face stays clear
MINIMAL ACCENT / TOP LEFTA tiny premium accent only. No decorative effects over your face. Hand gestures only change its color.
CONTROL THE EXPERIENCEPrivate by design · no video recording
02 / THE VISION UPGRADE
WHAT'S NEW IN V2
NV. / ENGINE 02

NOT JUST SEEN.
FELT.

One tiny premium accent sits in the top-left corner, with a clear unobstructed face mesh. Hand gestures update the accent color without launching extra effects.

01MINIMAL ACCENT
02TRACKING STREAMS
00FRAMES STORED BY APP
03 / BEHIND THE SIGNAL
HOW IT WORKS

EVERY MOVEMENT
MEANS SOMETHING.

Real-time landmarks become geometry. Your hand gestures change the color of one small top-left ornament without covering your face.

01 —

FACIAL GEOMETRY

MediaPipe Face Landmarker maps hundreds of points across the face, rendered as a luminous, reactive wireframe.

478 POINTS (MODEL-DEPENDENT)
02 —

GESTURE CONTROL

Show fingers to change the color of the tiny top-left accent. No other special effects are drawn on the stage.

HEURISTIC FINGER COUNT
03 —

REACTIVE LIGHT

One tiny accent stays in the top-left of the camera stage. The face remains clear, even when you move toward the corner.

VISION ENGINE 02 · CANVAS 2D
04 / THE TRUST LAYER
PRIVACY FIRST
YOUR CAMERA. YOUR CONTROL.

YOUR FACE ISN'T
OUR DATA.

We do not implement account creation, image recording, biometric identification, frame uploads, or a database. Your camera frames are processed by MediaPipe in your browser for this visualization; manual effect previews work without camera access. Camera permission is requested only after your action; Stop releases the video tracks, and leaving the tab stops the camera.

External connections: This version downloads MediaPipe code from jsDelivr and AI models from Google storage. Google's MediaPipe documentation notes that usage and performance metrics may be sent to Google. No software can guarantee absolute privacy or security; review the dependencies and host configuration before production use.