ABSTRACT
Person identification, in other words, biometrics, has been an open research area in recent years. Since computer-vision techniques have emerged, much has been done to automatically identify and recognize people. The main problem concerns the lack of security regarding written-text authentication (i.e., passwords), not being reliable anymore. Thus, an interesting approach using biometric-based information, such as face, fingerprint, and iris recognition, seems to be the new trend to distinguish different individuals. Although these techniques seem to be straight forward, one can consider them invasive methods. With respect to this, a noninvasive method can be formulated considering individuals handwriting styles. In this chapter, we are interested in identifying individuals, based on their handwriting patterns by means of machine-learning techniques, which can learn proper information from basic handwritten tasks, and those giving a probability of being an individual depending on the handwritten skills. Particularly, we are interested in deep learning techniques (i.e., convolutional neural networks) because of their ability into learning features in a hierarchical structure, similarly to the human learning procedure.
