Feature Extraction from Text Images to Study Individuality of Handwriting

Individuality of handwriting originates from interpersonal variations in handwriting. In this work, individuality of handwriting is studied using the following features: skew of writing, baseline, zone, slant and pen-pressure. Algorithms to extract the features are described and experimental results on controlled dataset collected from specific age groups are given. Experimental results are also given for sample images from IAM dataset. It is observed that for pre-defined text, skew-angle and baseline do not show much variation, whereas zone, slant and pen pressure show enough variation and hence ascertain individuality.