Abstract
The automatic processing of handwriting samples is part of the computational biometric. It applies qualitative and quantitative techniques by means of capturing, visualizing and analyzing handwriting. The main applications are writer identification and text understanding. Two significantly different situations appear: on line and off line data capturing. In the former the samples are obtained in a dedicated framework, where the writing instrument and the surface have several sensors. In the latter, the unique information available comes from the residues left on the paper. This chapter deals with the second situation. Width, grey value, direction and other parameters of the residual manuscript text are influenced by the psychomotor characteristics of the writer. Some of these personal parameters may be estimated from the observable properties of the written text.
TopBackground
The information recovered off-line from the static residues left on the paper is called pseudo-dynamic. It is clearly different from the actual dynamic information captured on-line during the writing activity. There are many contributions proposing different pseudo-dynamic characteristics, most of them oriented to signature authentication, which is a sub-problem of identification of the author of free text. Some of them are summarized in the next paragraphs.
Key Terms in this Chapter
Graphology: It is a pseudoscience that claims its ability to identify the writer, indicate his or her psychological state at the time of writing, or evaluate his or her personality by means of the analysis of the physical characteristics and patterns of handwriting. The term is sometimes incorrectly used to refer to forensic document examination.
Questioned Document Examination: Forensic phrase used to refer documents that are potentially disputed in a court of law. The primary purpose is to provide evidence about a suspicious or questionable document. Evidence might include alterations, chain of possession, damage, forgery, origin and authenticity, among aspects.
Authentication: The acceptance or the rejection of the identity given by a person.
Identification: The determination of the identity of a person, by a biometric system.
On-Line: Acquisition of biometric parameters in real time. In handwriting analysis, the writing instrument, the surface or both have sensors capturing position, trajectory, force and inclination.
Biometric Systems: Assemble of equipments, hardware and software whose purpose is to authenticate or identify persons automatically. They are based on the capture of several biometric parameters and the comparison of them with the same set of parameters already known of identified persons.
Off-Line: Acquisition of biometric parameter from the residues left by the human being actions. In handwriting analysis these residues are: the gray level and the wide of the strokes including the form of the stroke.
Co-Occurrence Matrix: A statistical matrix use for measuring texture in images. Could be considered an image by itself.
Biometry: The science that procures to obtain numeric value from biological phenomena and the use of such data in the automatic identification or authentication of human beings.
Pseudo-Dynamics Characteristics: Parameters acquired off-line used as estimators for on-line data.