An Adaptive Identification System Based on Chinese Handwriting

Automatic Chinese handwriting verification is a technique to online verify the handwriting with those in the pattern library by machines. This paper presents an adaptive online Chinese handwriting verification method for personal identification. A group of 4th order B-spline wavelet is developed to make wavelet decomposition for each stroke of the handwriting according to wavelet decomposition formulas of multi-resolution analysis. Furthermore, handwriting feature values of strokes are synthesized through a series of novel formulas so as to extract handwriting features which are rotation invariant, translation invariant and scale invariant. The properties of generalized Rayleigh quotient are analyzed and a stable and reliable pattern library is constructed, which can be dynamically adapted in runtime to model the variation of handwriting. The simulation results show that the proposed method has better stability and reliability.