Title: Fracture Mechanics-Based Quantitative Matching of Forensic Evidence Fragments

Abstract: Fractured metal fragments with rough and irregular surfaces are often found at crime scenes. Current forensic practice visually inspects the complex jagged trajectory of fractured surfaces to recognize a ``match'' using comparative microscopy and physical pattern analysis. We developed a novel computational framework, utilizing the basic concepts of fracture mechanics and statistical analysis to provide quantitative match analysis for match probability and error rates. The framework employs the statistics of fracture surfaces to become non-self-affine with unique roughness characteristics at relevant microscopic length scale, dictated by the intrinsic material resistance to fracture and its microstructure. At such a scale, which was found to be greater than two grain-size or micro-feature-size, we establish that the material intrinsic properties, microstructure, and exposure history to external forces on an evidence fragment have the premise of uniqueness, which quantitatively describes the microscopic features on the fracture surface for forensic comparisons. The methodology utilizes 3D spectral analysis of overlapping topological images of the fracture surface and classifies specimens with very high accuracy using statistical learning. Cross correlations of image-pairs in two frequency ranges are used to develop matrix variate statistical models for the distributions among matching and non-matching pairs of images, and provides a decision rule for identifying matches and determining error rates. A set of thirty eight different fracture surfaces of steel articles were correctly classified. The framework lays the foundations for forensic applications with quantitative statistical comparison across a broad range of fractured materials with diverse textures and mechanical properties.