Handwriting Analysis: Presented By: Dr. Alicia Carriquiry

Handwriting Analysis
Presented by: Dr. Alicia Carriquiry
Date forensicstats.org
CSAFE 1.0 Accomplishments - I
Project Title: G – Towards a Score-Based Likelihood Ratio for Handwriting Evaluation
Major Accomplishments:
• Collaborated with LAPD research on study of handwriting complexity
• Research paper on reliability of signature complexity assessments
• Draft manuscript on machine learning techniques for objective assessment of signature complexity
• Consult on design of study to relate complexity to handwriting expert analyses
• Developed handwriter software to analyze handwriting images and extract letters/features for use in
modeling handwriting evidence.
• Research manuscript being revised for Statistical Analysis and Data Mining
• Developed a hierarchical model for assessing questioned sample vs databased of potential authors. Model
outputs a posterior probability of writership for each potential writer in closed set.
• Two research manuscripts in preparation.
• Data collection of 900+ writing samples from over 100 participants to support research in this area.
• Article describing the data published in Data in Brief.
• Multiple invited presentations.
forensicstats.org | 2
handwriter
handwriter
Statistical modeling
Output
Results I
CSAFE 1.0 Accomplishments - II
Project Title: G – Towards a Score-Based Likelihood Ratio for Handwriting Evaluation
Impact:
• Open-source software to analyze handwriting and identify features is a major new resource for
the community. (Existing analogues are costly)
• Publicly available data is a significant new resource for the handwriting analysis research
community
• Novel methodological publications on the analysis of signature complexity assessments and on
analysis of handwriting samples (in a closed set)
Research Area Objectives
Research Area Objectives

Forensic document examination has seen the development of a number of automated systems
(e.g., FISH, WANDA, FLASH ID) capable of extracting features of writing that can then be
compared between documents.
-These systems are expensive and their algorithms are proprietary.
CSAFE has developed an open-source program called handwriter that outputs graphs, or
geometric representations of handwriting, that are reminiscent of the graphemes produced by
FLASH ID.
-Used to estimate the posterior probabilities of authorship of a questioned document for
writers in a closed set.
CSAFE 2.0 Objectives:

• Expanding and enhancing the statistical model to estimate probability of writership.
• Revising handwriter, to produce beta version.
• Continuing to collect standard writing samples from a large set of writers
CSAFE 2.0 Handwriting Projects and Lead Investigators
HW I- Handwriting Evaluation
Lead PI: Alicia Carriquiry, ISU
Co-PI: Danica Ommen, ISU
Team
Alexandra Arabio (Undergrad), Dr. Amy Crawford, Madeline Johnson (Grad), Dr. Lotem
Kaplan, James Kruse (Grad), John Libert (NIST), Madison McGregor (Undergrad), Robert
Ramotowski (NIST), Anyesha Ray (Undergrad), James Taylor (Undergrad).
Collaborators
Gary Licht (Iowa DCI).
CSAFE 2.0 – Research Plan
• Better understand limitations of current modeling approach,

including the number of clusters in which to group similar graphs.
• Explore utility of adding document-level and graph-level features to
probability model.
• Relax the requirement that the set of potential writers be closed.
• Extensively stress-test handwriter. Validate the program on a wide
variety of writing samples. Produce a well-documented beta version
to community of potential users.
• Design a quality metric for handwriting that can be included as part
of the estimate of probability of writership.
Resources and Needs
• There are at least three ways in which input and participation by

questioned documents examiners (QDEs) will be critical:
• Collection of realistic handwritten samples to add to the database.
• Beta-testing of handwriter.
• Calibration of handwriter relative to QDEs assessments.
Thanks for joining us today
We welcome your insights, suggestions, participation!
Please contact us:

www.forensicstats.org
Alicia@iastate.edu

Handwriting Analysis: Presented By: Dr. Alicia Carriquiry

Uploaded by

Copyright:

Available Formats

Handwriting Analysis: Presented By: Dr. Alicia Carriquiry

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Handwriting Analysis: Presented By: Dr. Alicia Carriquiry

Uploaded by

Copyright:

Available Formats

Handwriting Analysis

Presented by: Dr. Alicia Carriquiry

Project Title: G – Towards a Score-Based Likelihood Ratio for Handwriting Evaluation

Research Area Objectives

CSAFE 2.0 Objectives:

• Better understand limitations of current modeling approach,

• There are at least three ways in which input and participation by

We welcome your insights, suggestions, participation!

Please contact us:

You might also like