A Novel Technique for Unwarping Curved Handwritten Texts Using Mathematical Morphology and Piecewise Affine Transformation

Poor handwriting is a diagnostic criterion for developmental coordination disorder. It is typically characterized by its low overall quality and highly variable spatial organization of the letters. Detection of curved handwritten text is a continuing study in image analysis. It is characterized by various warped base-lines. The worst-case scenario of a poorly written text is often accompanied by arbitrarily warped and uncorrelated base-lines. The proposed methodology described in this paper provides a novel technique for offline detection and warp-correction of poorly written texts with heavily warped base-lines to improve its readability and facilitate its analysis. The algorithms utilize mathematical morphology accompanied by adaptive kernel management to obtain the basic line maps, and use out of phase warping to obtain the final image. Encouraging results were obtained by the use of this methodology, which was tested on different images specifically taken to represent the worst case scenarios.