Expert Assisting Computerized System for Evaluating the Degree of Certainty in 2D Shoeprints

• NCJ Number: 250336
• Author(s): Yoram Yekutieli; Yaron Shor; Sarena Wiesner; Tsadok Tsach
• Date Published: September 2016
• Length: 75 pages
• Grant Number(s): IAA-2009-DN-R-090
• Agencies: NIJ-Sponsored
• Publication Type: Research (Applied/Empirical); Report (Study/Research); Report (Grant Sponsored); Program/Project Description; Grants and Funding

Annotation

This project developed three tools for evaluating the rareness of randomly acquired characteristics (RACs) - such as scratches, nicks, tears, and holes - on shoe sole test impressions, as well as how to address the presence of an array of RACs on one sole.

Abstract

One tool developed is a large, consistent, and adequate database of RACs; all the RACs were scanned and their contours digitized. A second tool consisted of statistical models for determining the chances of finding similar RACs on other soles, as defined by shape, location, and orientation. A third tool involved practical techniques for a shoeprint expert to use in marking new RACs on other soles and quickly evaluating their evidential value in a more scientifically reliable manner than previous techniques. In order to create the database, test impressions were prepared from worn shoe soles. RACs appearing on the sole were marked by a qualified examiner using the specially designed "MarkAccidentals" software developed in this study. Each characteristic was automatically stored in the database with its digital representation, including its location and calculated orientation. Nearly 9,000 RACs were recorded and stored. The statistical algorithm developed enables the statistical evaluation of shoeprint RACs, which calculated the probability of finding another feature similar to a particular feature of a scanned and digitized RAC. The system can calculate the chance of getting another RAC in the same location or with the same orientation as the examined one. The number produced at the end of the process provides the expert with more objective and accurate indications of a match between a shoeprint and a particular shoe. 50 figures and 43 references