Handwriting and Hand Drawing Velocity Modeling by Superposing Beta Impulses and Continuous Training Component

We present in this paper a new strategy of handwriting or hand drawing velocity modeling Based on the Beta theory. The introduced approach aims to improve the interpretability of the dynamic profile model, reduce the data redundancy, and ameliorate the features accuracy. Indeed, we showed that the curvilinear velocity of handwritten or hand drawn trajectory can be rebuilt by superposing two components; consecutive Beta impulses representing its amplitude alternation imposed by the trajectory curvature variation and a velocity gain part of persistent pen carrying called "continuous training component" interpreting the learning level of the hand drawing faculty and the control of neuromuscular pulses synchronization. The proposed strategy was validated by the reduction of the error of curvilinear velocity fitting and the improvement of the recognition rate of Arabic handwriting characters represented by its model features vector.