What is it about?

Handwritten text gives away hints about the identity of a writer: Not all possible movements can be made by all possible hands. The combination of a taught writing style, the pen grip and the biomechanical characteristics of the hand and arm bones and muscles leads to individual patterns that can be described using pattern recognition. Although the reliability is less than for other biometric sources, such as the iris of the eye or DNA, handwriting is an example of behavior-based biometrics and has performances above 90% for finding one in 1000 writer candidates, if both the questioned document and the candidate consist of about a paragraph of text. It is risky to draw conclusions from smaller sample sizes. A good method addresses both the high-level letter (character) features as well as the low-level ink trace features.

Featured Image

Why is it important?

The identification of a writer can be of use both in a forensic and in a historic (paleographic) context. Who has written this document? Additional questions are: 'When' and 'Where' was the document produced. Worldwide, several projects are running concerning this problem. For example, for the Dead Sea Scrolls, it may be interesting to track a writer among the many fragments that were found.

Perspectives

Since this chapter was published, many more publications on writer identification and verifications have been published, world wide. It is interesting that handwriting still draws attention, although text input is often realized by keyboard-based methods, today. However, even today (2018), most students at a lecture have a paper notebook & pen with them. In any case, the historical handwritten texts are in the archives in massive amounts, waiting to be analyzed by computer-based methods.

prof. dr. Lambert Schomaker
Rijksuniversiteit Groningen

Read the Original

This page is a summary of: Writer Identification and Verification, January 2008, Springer Science + Business Media,
DOI: 10.1007/978-1-84628-921-7_13.
You can read the full text:

Read

Contributors

The following have contributed to this page