U.S. flag

An official website of the United States government, Department of Justice.

Quantitative Measures for Footwear Impression Comparisons

NCJ Number
303985
Author(s)
Martin Herman, Ph.D.; Steven Lund, Ph.D.; Hari Iyer, Ph.D.; G¨unay Do˘gan, Ph.D.; Adam Pintar, Ph.D.; Yooyoung Lee Ph.D.
Date Published
2021
Length
13 pages
Annotation

The goal of this project has been to research and develop quantitative methods for two-dimensional footwear evidence assessment to assist examiners by providing empirical support for their findings.

Abstract

Although research has produced many similarity metrics and scores for quantifying pattern comparisons, pattern evidence disciplines have long had difficulty in applying their methods to casework so that it withstands scientific scrutiny. This project shifts the focus from the question “What is the weight of evidence?” to “What relevant information is available to help assess the weight of evidence?” This project viewed this shift in focus as a significant advance for both the footwear impression discipline and the field of pattern evidence. The project’s goal is to develop an end-to-end system called Footwear Impression Comparison System (FICS). This system is based on a workflow that reduces the potential for bias by eliminating side-by-side human evaluation of the questioned impression and test impressions, parallels the analytic sequence examiners already follow, and provides quantitative support for examiner findings in harmony with the Scientific Working Group for Shoeprint and Tire Tread Evidence (SWGTREAD) conclusion scale. The first version of this end-to-end workflow was made into a proof-of-concept graphical user interface application, which is described in this report. Another paper details subsequent work to replace each component of the initial version, focusing on performance and casework utility. This report advises that “successfully completing the long-term research would be a transformative landmark in pattern evidence evaluation.” 2 figures and a listing of 5 research papers, 2 talks, and an online software from this project

 

 

Date Published: January 1, 2021