The statistical properties and coding of handwriting and hand-drawn graphics in tutorial classes

Guifo Guifo, Robinson (1989). The statistical properties and coding of handwriting and hand-drawn graphics in tutorial classes. PhD thesis The Open University.

DOI: https://doi.org/10.21954/ou.ro.0000d3a5

Abstract

This research investigates compression techniques as applied to input from a digitizing tablet. The representations we consider are either drawn from the literature or developed within this thesis; they involve selecting relevant points along the trajectory of the writing pen, and using an interpolating function for reconstruction. The two dimensional parametric polynomial interpolator (e.g linear, quadratic, cubic ) is used for regenerating the trajectory of the pen. Several techniques for selecting relevant points , i.e control points are discussed. We determine our best representation by weighting the criteria, analyzing the experimental results, examining the theoretical considerations and, making where necessary, justifiable tradeoffs. The techniques tested are objectively evaluated in terms of accuracy, efficiency and compactness.

The information characteristics (i.e signal entropy) of the handwriting and drawing signals, are estimated from the original and approximated data. Results are discussed and conclusions drawn. The research has been carried out on static data.