Related Experiment Videos
Individuality of handwriting
Sargur N Srihari1, Sung-Hyuk Cha, Hina Arora
1Departmentof Computer Science and engineering, Center of Excellence for Document Analysis and Recognition, University at Buffalo, State University of New York, 14228-2567, USA. Srihari@cedar.buffalo.edu
Journal of Forensic Sciences
|July 26, 2002
Summary
Handwriting is unique to each individual. This study used computer analysis of 1,500 writing samples to objectively prove handwriting individuality, supporting its use as evidence in court.
Area of Science:
- Forensic Science
- Biometrics
- Computer Science
Background:
- Court rulings question the scientific validity of handwriting analysis.
- Forensic document examiners (FDEs) rely on subjective assessments of handwriting individuality.
- Objective validation of handwriting uniqueness is needed for legal admissibility.
Purpose of the Study:
- To objectively validate the hypothesis that handwriting is individual.
- To develop a quantitative method for assessing handwriting uniqueness.
- To provide scientific support for the admissibility of handwriting evidence in legal proceedings.
Main Methods:
- Collected handwriting samples from 1,500 diverse individuals.
- Utilized computer algorithms to extract handwriting features from scanned images.
- Applied machine learning approaches to quantitatively establish writer individuality.
Main Results:
- Identified characteristic handwriting attributes like line separation, slant, and character shapes.
- Demonstrated high-confidence writer identification using global handwriting features and minimal character data.
- Achieved objective validation of handwriting individuality.
Conclusions:
- Handwriting possesses unique, quantifiable characteristics.
- Computational analysis can reliably determine writer identity.
- This research supports the scientific basis for admitting handwriting evidence in court and aids forensic document examiners.