Identification and decomposition of fast handwriting system

Systems for carrying out automatic handwriting recognition or writer verification based on real-time handwriting signals have as their basic function the separation and characterization of a muscular force function and a hand system impulse response. The problem is how to deconvolve the handwriting signal into two functions mentioned above. The problem assumes that the only datum available is the convolution itself; the handwriting signal. Limiting our discussions exclusively to the "fast" handwriting case, an approach to this class of blind deconvolution -is described based on the interaction of theory and experiments.