Morphing robust face recognition
Type: Bachelor EE/CS
If you are interested please contact :
Face morphing is a technique to blend facial images of two or more subjects such that the result resembles both subjects. Face morphing attacks pose a serious risk for any face recognition system. Without automated morphing detection, state of the art face recognition systems are extremely vulnerable to morphing attacks. The most common approach to face morphing is by rst detecting landmarks in both contributing faces, then dene triangles in the images, determine an average geometry and map the averages of the textures in the triangles from the contributing faces to the averaged geometry.
Figure 1: From left to right: landmarks on face 1 and 2; triangulated geometry; morphed face.
The aim of this assignment is to investigate morphing robust face recognition. The proposed approach is to first create a training set with geometrically normalised images. First a reference geometry is dened e.g. based on landmarks of "the average face". Next the texture of all faces is mapped on this reference geometry, resulting in a set of faces with the same geometry, but different textures. Using this trainingset a face recognition method is trained. If the resulting face recognition system compares two faces to determine if they match, the faces first have to be geometrically normalised. We expect that the resulting face recogntion system is more robust against morphing attacks. An interesting question is also how well the resulting face recognition system performs in the task of face recognition.