Statistical Approaches to Type Determination of the Ejector Marks on Cartridge Cases. Issue 2 (2nd May 2017)

• Record Type: Journal Article
• Authors: Warren, Eric M.; Sheets, H. David
• Is Part Of: Journal of forensic sciences. Volume 63:Issue 2(2018)
• Page Start: 431
• Page End: 439
• Publication Date: 2017-05-02
• Subjects: forensic science -- Glock -- Smith & Wesson Sigma -- ejector marks -- morphometrics -- Naïve Bayes classification; Medical jurisprudence -- Periodicals; Forensic sciences -- Periodicals; Forensic Medicine -- Periodicals; Gerechtelijke geneeskunde; Gerechtelijke chemie; Gerechtelijke psychiatrie; 363.2505
• DOI: 10.1111/1556-4029.13529
• Languages: English
• ISSNs: 0022-1198

Abstract

Abstract: While type determination on bullets has been performed for over a century, type determination on cartridge cases is often overlooked. Presented here is an example of type determination of ejector marks on cartridge cases from Glock and Smith & Wesson Sigma series pistols using Naïve Bayes and Random Forest classification methods. The shapes of ejector marks were captured from images of test‐fired cartridge cases and subjected to multivariate analysis. Naïve Bayes and Random Forest methods were used to assign the ejector shapes to the correct class of firearm with success rates as high as 98%. This method is easily implemented with equipment already available in crime laboratories and can serve as an investigative lead in the form of a list of firearms that could have fired the evidence. Paired with the FBI's General Rifling Characteristics (GRC) database, this could be an invaluable resource for firearm evidence at crime scenes.