Book Title: Computational Criminology: AI Applications in Forensic Science and Criminal Justice
Editors: Dr. Xavier Louis, Dr. Surbhi Girdhar, Ms. Aswathi Chandran Nair, Mr. Ravi Kumar, and Ms. Nandini Katare
Chapter: 12
DOI: https://doi.org/10.59646/704/12
Author: Sneha Nair
Abstract
Firearm and toolmark identification have occupied an uneasy position in forensic science for nearly a century. Calvin Goddard’s 1926 work established the basic premise that the markings left on a fired bullet or cartridge case by the working surfaces of a firearm contain individualising features that can, in principle, be matched between a questioned exhibit and a known firearm.[1] Alfred Biasotti’s 1959 study attempted to put this premise on quantitative footing by counting matching consecutive striations on bullets fired from the same and different firearms, anticipating the statistical approach that would eventually become standard.[2] For most of the intervening period, however, the discipline relied on expert pattern comparison and categorical conclusion language (“identification”, “elimination”, “inconclusive”) that conveyed certainty the underlying methodology did not, on contemporary scientific standards, justify.[3]
The discipline’s methodological challenge has accordingly been twofold. The first task is to develop quantitative similarity measures between toolmark images that can be statistically validated, replacing the categorical pattern judgments of the traditional method with calibrated probabilistic conclusions. The second is to integrate these measures into an evidentiary framework that satisfies contemporary admissibility standards error rates, validation studies, replicable procedures while preserving the operational utility of the comparison for investigators and courts.[4] Computational ballistics, and in particular the application of convolutional neural networks to two- and three-dimensional surface imagery of fired components, has emerged as the methodological response to both tasks.[5]
[1]Calvin H Goddard, “Scientific Identification of Firearms and Bullets” (1926) 17 Journal of Criminal Law and Criminology 254, 257.
[2]Alfred A Biasotti, “A Statistical Study of the Individual Characteristics of Fired Bullets” (1959) 4 Journal of Forensic Sciences 34, 36.
[3]National Research Council, Ballistic Imaging (National Academies Press, 2008) 18.
[4]James E Hamby and James W Thorpe, “The History of Firearm and Toolmark Identification” (1999) 30 AFTE Journal 266, 268.
[5]Sinha, A. K. (2026). Whistleblower Protection and Criminal Investigations in Corporations. Minnesota Journal of Business Law and Entrepreneurship (1) 452, 454.