Writer Recognition by Analyzing Handwritten Documents and by Using Neural Network Classifiers.

Date of Submission

December 2007

Date of Award

Winter 12-12-2008

Institute Name (Publisher)

Indian Statistical Institute

Document Type

Master's Dissertation

Degree Name

Master of Technology

Subject Name

Computer Science

Department

Electronics and Communication Sciences Unit (ECSU-Kolkata)

Supervisor

Chanda, Bhabatosh (ECSU-Kolkata; ISI)

Abstract (Summary of the Work)

For many centuries handwriting has been used to identify an individual. The hypothesis about handwriting being a personal biometric relies on the fact the process of handwriting is an unconscious act learnt over time, and some pen movements are invariant and not easily changed when an attempt at forgery or disguise is made. When doubts of authenticity of handwriting arise, forensic document examiners are asked to conduct an analysis of the questioned document. They seek characteristics of handwriting (features) that are consistent in a person’s normal writing by analysis of shapes and structure of the handwriting. While other branches of forensic science, such as DNA analysis, have been explained and proven by experimental analysis, the forensic techniques used in handwriting analysis have far less scientific support. The methods used by the examiners are intutively reasonable and have been derived from experience. It is the credibility of the document examiner that has been a key basis in a court of law rather than scientific basis of the techniques.In recent cases the scientific acceptability of forensic analysis of handwriting has been successfully challenged. To provide a scientific support for the handwriting analysis pattern recognition techniques have been employed. It has been demonstrated that handwriting indeed can be used to identify an individual with high accuracy. However, it has not been shown that writers can be distinguised using the techniques forensic document examiner use in their analysis.A writer can be characterized by his own handwriting. The problem of writer identification arises frequently in the court of justice where one must come to a conclusion about the authenticity of a document (e.g. a will). It also arises in banks for signature verification. In order to come to a conclusion about the identity of an unknown writer, two tasks must be considered:1. The writer identi fication, where handwritten samples are retrieved from a database and the query sample is taken to find the writer of the query document.2. The writer veri fication, where two samples of handwriting are taken and a conclusion is drawn whether they are written by the same writer or notPrevious work: Most work in the writer recognition field has concentrated on signature verification, since signatures typically present more individuality. However, in many cases signatures are not available for analysis, only words or characters. Word based analysis work started with Steinke et al. [4]. Zois et al. [5] used features obtained from the morphological transformation of thinned word images to answer the different writer/same writer question using one single word Overview of Study: In case of handwriting two types of variation can occur: within writer variation and between writer variation. The first type of variation occurs within a writer’s hanwritten document. The second type of variation occurs between handwritten documents of two different writers. The goal of our study was to establish the intuition that within writer variance is less than the between writer variance . The study have four phases: data collection, preprocessing, feature extraction, and recognition . The first three phases was done by Tripathy[11]. In the recognition phase he used KNN classifier, whereas I have used Neural Classifier. In the data collection phase handwriting samples of different writers were colllected. Preprocessing was done on the images so that the suitable features can be extracted. This phase have three steps: noise removal, getting line count and extracting words boundaries. In feature extraction phase required features of each word of a document were extracted from the preprocessed image for discriminating two different writing style. Recognition phase helps to prove that within writer variance is less than between writer variance.

Comments

ProQuest Collection ID: http://gateway.proquest.com/openurl?url_ver=Z39.88-2004&rft_val_fmt=info:ofi/fmt:kev:mtx:dissertation&res_dat=xri:pqm&rft_dat=xri:pqdiss:28843074

Control Number

ISI-DISS-2007-206

Creative Commons License

Creative Commons Attribution 4.0 International License
This work is licensed under a Creative Commons Attribution 4.0 International License.

DOI

http://dspace.isical.ac.in:8080/jspui/handle/10263/6369

This document is currently not available here.

Share

COinS