Studying the variability of handwriting patterns using the Kinematic Theory

Area of Science:

• Motor Control
• Human-Computer Interaction
• Biomechanical Engineering

Background:

• Handwriting recognition systems in PDAs and smartphones can be improved by integrating motor control knowledge.
• Variability in handwriting stems from neuromuscular system (NMS) and central nervous system (CNS) task scheduling, manifesting as global and local fluctuations.
• The Kinematic Theory of Rapid Human Movement provides a framework for analyzing these variations.

Purpose of the Study:

• To analyze handwriting variability using the Sigma-Lognormal model.
• To generate artificial handwriting distortions to study scale changes and rotational deformations.
• To provide practical insights for developing robust handwriting recognition systems.

Main Methods:

• Utilized the Sigma-Lognormal model, a key component of the Kinematic Theory of Rapid Human Movement.
• Artificially generated handwriting distortions by manipulating Sigma-Lognormal parameters.
• Conducted an ANOVA analysis on handwriting data from six writers.

Main Results:

• Generated a wide range of handwriting distortions, revealing patterns related to scale changes and rotational deformations.
• Experimental results corroborated theoretical predictions, validating the Kinematic Theory's relevance.
• Findings align with previous studies on single strokes using the Sigma-Lognormal model.

Conclusions:

• The study provides a deeper understanding of motor control in handwriting.
• Results offer practical guidance for creating extensive databases for training and testing handwriting classifiers.
• The Sigma-Lognormal model is effective for analyzing and synthesizing handwriting disruptions, enhancing recognizer design.