Face Morpher

Seamless affine transformation the facial features from one face to another.

Face Morpher seamlessly morphs the facial features from one face to another by stretching and squishing them to match. This is accomplished by identifying facial landmarks using dlib’s 68-point landmark detector and creating triangles using the Delaunay triangulation algorithm. These triangles are then subjected to affine transformation, ensuring alignment with the features of the target face. A video can be generated using linear interpolation, where the transformation is applied iteratively. You can find the python implementation here.