Feature Enhancement and Denoising of a Forensic Shoeprint Dataset for Tracking Wear-And-Tear Effects
Xavier Francis, Hamid Sharifzadeh, Angus Newton, Nilufar Baghaei, Soheil Varastehpour · 2019
Shoeprints found at crime scenes are valuable sources of information for crime scene investigators. Forensic scientists and law enforcement maintain databases of reference shoeprints to aid in the identification of crime scene prints. Reference shoeprints are taken by experts in a laboratory setting and go through a pre-processing stage before being entered into the database. On a new and unique dataset of shoeprints collected for the analysis of wear-and-tear effects, we describe a shoeprint enhancement method that allows for undesirable features to be mitigated while maintaining valuable features that capture the effects of wear-and-tear. Based on objective evaluations the method performs well.