Pose compensation of facial images using 3D morphable models for improved face recognition
Type: Master EE
Duration : April 2017 until unknown
If you are interested, please contact:
Face recognition works well for frontal facial images, but degenerates quickly for images under pose. Morphable models are powerful 3D models that can be used to estimate the 3D pose of a face from a single image. In addation they can be used to synthesize a frontal image of the face. The idea is to use the synthesized frontal face image to perform facial recognition on and in this way obtain better face recognition performance for face images with pose.