Abstract
A development of an on-line behavioral automatic script verification system is proposed for verifying the identity of a person. The proposed system is based on the dynamic movement of pen during hand writing for each script, computing a number of features that are scale and displacement
invariant. For each script, the static features like the normalized central moments (NCM) and the normalized Fourier descriptors (NFD) are compared to the dynamic features like the speed of writing individual scripts. In both cases the data acquisition, pre-processing, feature extraction and
comparison steps are analyzed and discussed. Simulations comparing the performance of each system are considered.