|Topic:||3D Face recognition|
|Supervisor:||Ing. Vojtěch Franc Ph.D.|
|Description:||The current state-of-the-art methods for face recognition are based on black-box Convolutional Neural Networks with no explicit knowledge about the face. As a result training the CNN requires large amounts of annotated facial examples which have to cover all the image variability. The goal of the thesis will be to combine CNNs and 3D face model in order to improve generalization ability of the prediction system and at the same time to reduce the number of annotated training examples. The developed prediction system will be evaluated and compared against existing methods on standard face recognition tasks like gender, age and identity estimation.
|Bibliography:||- L. Tran and X. Liu. Nonlinear 3D Face Morphable Model. CVPR 2018.
- V. Blanz and T. Vetter. Face recognition based on fitting a 3D morphable model. PAMI, 2003.
- V. Blanz and T. Vetter. A morphable model for the synthesis of 3D faces. In Proc. of the conference on Computer graphics and interactive techniques. 1999.