Handwriting (Question Document Analysis)
Forensic document examiners perform a range of different tasks. These can include analyses of writing, papers, inks, etc. CSAFE activity has focused on two topics in handwriting: (1) Examiners are often asked to assess a questioned signature against samples from a known writer to determine if the questioned signature is genuine or simulated; (2) examiners are asked to examine a document or writing sample and attempt to match it to a known writer (or population of known writers available in the database). These are described briefly below.
- Providing statistical support for a study of signature complexity (subjective assessment of complexity, relating complexity to objective dynamic [taken while the person is signing] measures of the writing, determining examiner performance as a function of complexity).
- Developing statistical methods for measuring the complexity of a static signature sample. Note that a method for static samples is required to make this approach practical.
- Developing open-source software to identify relevant features (e.g., loops, crossings) in a handwriting sample.
- Developing statistical approaches that use counts of relevant features to compare a questioned handwriting sample with a collection of writing samples from known authors.
Benefits of Research
The two CSAFE projects in this area are showing how statistical methods can enhance the work of forensic document examiners. The signature project attempts to rigorously assess the role of complexity in signature analysis and relate complexity to examiner performance. The results of the study will be of great use to examiners and to the legal system. The handwriting project is attempting to develop open-source software and publicly available statistical algorithms for writing comparison. This is an important topic as document examiners try to better integrate quantitative approaches in their work.
Select Publications, Conference Papers, Presentations and/or Tools
“Assessing the Complexity of Handwritten Signatures”, Stern, Angel, Cavanaugh, Lai, Zhu, to appear in Law, Probability and Risk
Poster “Statistical Analysis of Letter Importance for Document Examination” presented at the American Academy of Forensic Sciences Annual Scientific Meeting, February 2018.
Presentation, International Conference on Forensic Inference, Minneapolis, MN, September 2017.