Segmentation and analysis of handwritten scripts from patients with neurological diseases
Elena Kalcheva-Yovkova, Georgi Gluchev · 2003
A method for automatic segmentation and analysis of handwritten scripts belonging to patients with ND and healthy individuals is proposed A combination of connected component analysis and projection profile technique is used to separate the text-lines. Several common geometric features describing size. orientation, shape and baseline deviations of the handwriting are computed and statistically examined. The baseline variability and the aspect raho revealed differences between ND-patients and control group. Also the changes in the stroke's width and height in the beginning and the end of the Writing process can be used to separate the clinical Parkinsonians from the others. Nevertheless additional features and data are required to obtain reliable evaluation of the handwriting changes due to neurological diseases.