Ralph Niels & Louis Vuurpijl Nijmegen Institute for Cognition and Information Radboud University Nijmegen The Netherlands
Handwriting styles and copybooks Method Results Discussion
Handwriting is individual Similar handwritings: handwriting styles Top down ‘copybooks’ * We defined writing styles bottom up * S.-H. Cha, S. Yoon, C.C. Tappert, 2006.
Handwriting recognition Personalized recognizers Handwriting synthesis ‘Handwriting fonts’ Forensic writer identification Human experts use the notion of style
Databases: Unipen trainset Unipen devset Plucoll database Online handwritten characters (pre-segmented) 43 writers 41 writers
The prototype we used are averaged shapes of actual handwritten characters L. Vuurpijl & L. Schomaker, Finding Structure in Diversity, ICDAR R. Niels, L. Vuurpijl & L. Schomaker, Automatic allograph matching in forensic writer identification, IJPRAI, Feb
PC j PC k PC i Prototypes Prototype clusters
Relative frequency of the occurrence of each prototype cluster in a persons handwriting
PC j PC k PC i Prototypes Prototype clusters
Handwriting Z Handwriting X Handwriting Y
Hierarchical clustering of membership vectors (handwritings) HZXEDIAGBBJKYCF Handwriting Writing styles
PC j PC k PC i HZXEDIAGBBJKYCF
Monte Carlo simulation of combinations of parameters and levels Large number of writing styles Find the writing styles that occur most By prototypes or By writers
Copybooks Preliminary results Visual evaluation by handwriting experts Meaningful names Well-known broad categories: cursive, mixed and print
Mixed Cursive Print
Mixed Cursive Print
Applied to/with, not limited to: Online Latin characters Dynamic Time Warping for character comparison (human congruous) Best of both worlds: Integrate top down and bottom up (with forensic experts) HZXEDIAGBBJKYCF Integrate