3D Face Reconstruction
Face reconstruction · Face image or webcam capture
About 3D Face Reconstruction
This Space uses a MediaPipe-based face detector to extract facial landmark geometry. It displays the estimated points in a rotatable 3D plot and provides an OBJ export through its interface.
A clear, front-facing image with good lighting is a useful starting point. The result describes estimated face geometry; it is not a full textured head scan with reconstructed hair, ears, and hidden surfaces. Webcam access is optional and depends on browser permission.
How to use 3D Face Reconstruction
- Select Upload for an image, or use Webcam if you want to capture a frame.
- Keep the face visible and select Generate 3D Model.
- Check the status and 3D preview, then use the OBJ download if generation succeeds.
What to know before you start
- Uploads and captured images are processed by the third-party Space. Use an image you have permission to submit.
- Webcam use requires browser camera permission. Use Upload if camera access is unavailable.
The provider controls availability, sign-in, processing, and downloads. Save any result before reloading or closing the demo.
Does the demo reconstruct a complete textured head?
The inspected workflow exports detected face geometry. It does not show a full-head texture reconstruction stage.
