An algorithm for forensic toolmark comparisons

arXiv:2312.00032v2 Announce Type: replace-cross
Abstract: Forensic toolmark analysis traditionally relies on subjective human judgment, leading to inconsistencies and inaccuracies. The multitude of variables, including angles and directions of mark generation, further complicates comparisons. To address this, we introduce a novel approach leveraging 3D data capturing toolmarks from various angles and directions. Through algorithmic training, we objectively compare toolmark signals, revealing clustering by tool rather than angle or direction. Our method utilizes similarity matrices and density plots to establish thresholds for …

abstract algorithm analysis arxiv cs.cr cs.lg data human judgment mark novel stat.ap through training type variables