FacialDetection

Only in Wrap4D

Detects lip and eyelid contours on a face in 3D space using a generic detector based on machine learning.

Tip

The results of the detector are usually not very accurate. It is highly recommended to train a personalized detector in Track instead of using this node.

---

Editor

FacialDetection has an editor which display detection result on a model rendered through the camera

Inputs

Geometry

Geometry Geometry that is used for detection. The model should have texture

Camera

Camera Camera to render the geometry in 2D space to be used by the detector

Output

FacialDetection Facial detection

Parameters

Use Camera Resolution:

if set, the camera resolution is used

Camera Resolution:

render width and height that is passed to the detector

Auto Compute:

if set, the node will be recomputed each time a parameter or an input data is changed. This parameter is ignored in batch mode

Compute:

runs the detection

Export:

saves computed facial detection on disk in JSON format