Abstract
The tracking and monitoring of fugitives and persons of interest is of significant concern for the Indiana Department of Corrections (IDOC) Fugitive Detection Unit. The research conducted was to help determine the benefits of
implementing a face recognition technology solution. Images were analyzed for standard compliance to help determine their suitability for input into a face recognition matcher. Results from this analysis showed the images were not in compliance with the NIST Mug Shot Best Practices, nor could the software optimize the images to make them compliant. A visit to the intake facility indicated that the process by which these mug shots were collected needed to be addressed before face recognition technology could be implemented.
Consequently, the IDOC main prisoner intake facility’s current mug shot image capture process was assessed. Using the analysis from the images, along with observations from the mug shot capture process, an optimized capture process
was implemented for a trial period of two weeks to determine its effectiveness. Results show that the capture process improved the standard compliance of the mug shot images, determining that the images collected would be usable with
face recognition technology. Another finding was that the centerline location ratio variable, which has a precise threshold, was not compliant for any images in either dataset leading to the need for further study to determine if this variable should utilize a range of values for an operational environment such as at the IDOC.
