Development of face similarity linkage for the attribution of intelligence links in unsolved sexually motivated serial homicide

Abstract

Sexually motivated crime (SMC) has an immense and ongoing impact on victims, victim-survivors, their families, and the broader society, such that countless resources and forensic techniques have been deployed in pursuit of prosecuting perpetrators. For law enforcement investigating sexual crimes, linkage between incidents is an important tool to identifying serial offenders as well as victims that are unattributed or part of unsolved cold cases. Serial sex offending is well reported making it likely that operating or dormant sexually motivated offenders that have gone undetected and are still at large. To assist in solving these challenges a technique that can link victims of these crimes beyond victimology and modus operandi is needed. This research develops Face Similarity Linkage as a forensic tool to identify minute variations in facial anthropometry to detect similarities between victims of SMC as intelligence for further investigation. Several tests, detecting interobserver error and camera angle variation were used to determine which facial landmark measurements were most useful in differentiating three random, unrelated faces from an available dataset. Our study revealed 55 ratios of various point-to-point measurements that show promise for future studies using this technique. While there was some variation between readers, it was approximately 5% or less and could be improved with additional training of readers.