Method for aligning face image of person e.g. for detection of face features, involves training appearance model component with training data to estimate score function and to minimize angle between gradient- and ideal way directions

摘要

The method involves acquiring a face image of a person, and aligning a generic lattice face in the face image using a discriminative face alignment model that comprises generative form and discriminative appearance model components. The appearance model component is trained with a training data to estimate a score function that is formed as a function of a form parameter (28) of a given image and to minimize an angle (36) between a gradient direction (32) of the score function and an ideal way direction of alignment for the parameter, where the way direction represents a vector (34). Independent claims are also included for the following: (1) a system for aligning a face image of a person (2) a product comprising a set of instructions for executing a method for aligning a face image of a person.