Download full paper: Teulings_Schomaker_ActaPsychologica1993.pdf
Acta Psychologica 82 (1993) 69-88
Invariant properties between stroke features in handwriting
Hans-Leo Teulings and Lambert R.B. Schomaker
NICI, University of Nijmegen, Nijmegen, The Netherlands
A handwriting pattern is considered as a sequence of ballistic strokes. Replications of a pattern may be generated from a single, higher-level memory representation, acting as a motor program. Therefore, those stroke features which show the most invariant pattern are probably related to the parameters of the higher-level representation, whereas the more noisy features are probably related to the parameters derived at the lower levels (top-down hierarchy). This hierarchy of invariances can be revealed by the signal-to-noise ratio (SNR), the between-parameter correlations, and the between-condition correlations. Similarly, at the higher level a sequence of strokes may act as a unit from which individual strokes are derived (sequence hierarchy). This hierarchy of invariances can be revealed by the between-stroke correlation, which forms a weaker criterion than rescalability, which has been rejected mostly. Previous research showed that vertical stroke size has higher SNRs and higher between-condition correlations than stroke duration or peak force, whereas the latter two features were also negatively correlated. This suggested that vertical stroke size is a higher-level parameter than the other two. The present research largely confirmed this top-down hierarchy and even for upstrokes and downstrokes separately. Downstrokes were more invariant than upstrokes in terms of vertical stroke size. However, contrary to the vertical stroke size, the horizontal stroke size was not invariant. Both vertical and horizontal sizes showed substantial between-stroke correlations. In contrast, the stroke durations did not show any between-stroke correlations. This suggests that stroke segmentation is reliable in spite of the discrete sampling of the handwriting movements.

Download full paper: Teulings1994Impedovo.pdf
Teulings, H.L. (1994). Invariant handwriting features useful in cursive script recognition. In S. Impedovo (Ed.), Fundamentals of handwriting recognition (pp. 178-189). Berlin: Springer.
Invariant Handwriting Features Useful in Cursive-Script Recognition
Hans-Leo Teulings
Nijmegen Institute for Cognition and Information (NICI), P.O. Box 9104, 6500 HE Nijmegen, The Netherlands.
Abstract. The within-writer variability of handwriting forms one of the problemsin the automatic recognition of cursive script. Variability can be handled by choosing handwriting features based upon the process of handwriting generation or upon computational models. Handwriting patterns are represented by a sequence of motor actions, i.e., "strokes", which can be identified by invariant segmentation. Each stroke is characterized by features related to motor memory parameters which can be identified by their high signal-to-noise ratios.
Keywords. Cursive-script recognition, on-line handwriting, movement production models, computational models, movement variability, segmentation, movement features, signal-to-noise ratio, off-line reconstruction from on-line.

Earlier Publications

Teulings, H.L., & Schomaker, L.R.B. (1991). Unsupervised learning of prototype allographs in cursive-script recognition using invariant handwriting features. 2nd International Workshop on Frontiers in Handwriting Recognition (pp. 45-55). Bonas, France, September 24-27, 1991.

Teulings, H.L., & Schomaker, L.R.B. (1991). Invariant properties of handwriting motor programs to be employed in automatic cursive-script recognition. In G.E. Stelmach (Ed.), Proceedings of the 5th Handwriting Conference of the IGS. Motor control of handwriting (pp. 21-23). Tempe AZ: Arizona State University.

Teulings, H.L., Thomassen, A.J.W.M., & Van Galen, G.P. (1986). Invariants in handwriting: The information contained in a motor program. In H.S.R. Kao, G.P. Van Galen, & R. Hoosain (Eds.), Graphonomics: Contemporary research in handwriting (pp. 305-315). Amsterdam: North-Holland.