Face projector with blendshapes deformation and landmarks feedback [Graduation project module]
Developed as a part of the GP NoCAP to make markerless motion capture.
The module’s purpose is to load a 3D face model with its blendshapes, deform the face using the blendshapes’ values, project this face on a 2D plane and finally return the facial landmarks’ position after projection.
The module is written with C++ and OpenGL and the frame takes an average of 1 ms to calculate.