Dr Kate Seigfried- Spellar

Digital evidence is estimated to be a factor in about 90% of criminal cases, and mobile phones are the most seized devices. The popularity of smartphones has also increased the ease of sharing high-quality multimedia via social media and native texting. In CSEA criminal cases, images and videos can serve as important forms of digital evidence. As a part of the grooming process, CSEA offenders will send sexually explicit "sexts" to the minor, and although the offenders may try to conceal their faces, research has shown the validity of knuckles and fingernail beds as another form of unique biometrics. This presentation will discuss the research and development of the Suspect Hand Biometrics Tool - a digital forensics tool that automatically processes images and videos to extract knuckle and fingernail bed biometrics. To improve match performance, we have implemented additional preprocessing methods (e.g., segmentation) and evaluated several state-of-the-art recognition models. These enhancements yielded meaningful improvements in match performance, demonstrating dorsal-hand (i.e., back of the hand) recognition even in operationally realistic, unconstrained imagery. Finally, this presentation will include recommendations for collecting comparison and exemplar imagery for law enforcement investigations and provide use case scenarios demonstrating the Suspect Hand Biometrics Tool, which will be a free tool for law enforcement, funded by Safe Online.