Since face alignment seeks to deform a face model to match it with the features of the image of a face by optimizing an appropriate cost function, the authors of this paper propose a new face model that is aligned by maximizing a score function, which is learned from training data, and that is imposed to be concave.
The authors show that this problem can be reduced to learning a classifier that is able to say whether by switching from one alignment to a new one, the model is approaching the correct fitting. This relates to the ranking problem where a number of instances need to be ordered. For training the model, the authors propose to extend GentleBoost [23] to rank-learning. Extensive experimentation shows the superiority of this approach to other learning paradigms, and demonstrates that this model exceeds the alignment performance of the state-of-the-art. (Published abstract provided)