The study of identification algorithm based on palmprint algebraic features

This paper uses the proposed two-stage kernel fisher discriminate analysis method to extract algebra feature of palmprint. It is to take each piece of palmprint image as a point of a high dimensional space. The palmprint images from the training set form a training data set. Through a nonlinear mapping the input space of training data is mapped to a feature space, making different types of palm print data in feature space become linearly separable.