Footwear Databasing to Support Forensic Intelligence

The United States currently lacks a national footwear reference collection.  This limitation undermines the use of footwear and footwear impressions in criminal investigations.  With regard to intelligence, the lack of a footwear database makes it challenging to generate investigative leads, and to link scenes together based on repeated impressions.

The objective of this project is three-fold.  First, design an easy-to-use mobile application that can be employed without the need for specialized training.  The purpose of this application is to allow users to upload three images and associated metadata, including a footwear outsole, the shoe’s upper, and the manufacturer’s tag or label (if present).  The application design will be beta-tested, and adjusted before release, ensuring its ease-of-use by law enforcement personnel in custody suites when tasked with collecting arrestee records, as well as widespread general circulation to achieve the second goal of creating an opportunistic crowdsourced reference dataset.

To generate the aforementioned reference collection, the third objective must be realized, which is to create and evaluate a social media campaign to motivate citizen scientists to participate in the collection of forensic reference data.  The reference collection that will result from the citizen scientist aspect of this project will support several outcomes.  This includes the immediate need for a dataset that can be consumed by existing software/hardware intelligence solution applications, a benchmark for testing/evaluating proposed solutions, a community-dataset for make/model classification, and a resource for training examiners on image encoding tasks (and/or training neural nets for the same purpose).

In addition to the convenience sample of outsole images that will result and support the forensic footwear community, this project will evaluate the efficiency, suitability, and quality of the crowdsourced data.  The final deliverable will be best-practice recommendations that are sufficiently generalizable to other forensic disciplines, thereby providing a framework for future citizen science sourced projects related to forensic science and the criminal justice system.

Note: This project contains a research and/or development component, as defined in applicable law," and complies with Part 200 Uniform Requirements - 2 CFR  200.210(a)(14). CA/NCF