Firearms identification by the acoustic signals of their mechanisms. (January 2020)

• Record Type: Journal Article
• Authors: Giverts, Pavel; Sofer, Saad; Solewicz, Yosef; Varer, Boris
• Is Part Of: Forensic science international. Volume 306(2020)
• Publication Date: 2020-01
• Subjects: Classification -- Sound -- Acoustic -- Signal analysis -- Machine learning -- LDA -- XGBoost; Medical jurisprudence -- Periodicals; Chemistry, Forensic -- Periodicals; Forensic Medicine -- Periodicals; Médecine légale -- Périodiques; Chimie légale -- Périodiques; Gerechtelijke geneeskunde; Gerechtelijke chemie; Gerechtelijke psychiatrie; Chemistry, Forensic; Medical jurisprudence; Electronic journals; Periodicals; Electronic journals; 614.1
• DOI: 10.1016/j.forsciint.2019.110099
• Languages: English
• ISSNs: 0379-0738

Abstract

Highlights: Definition of acoustic signals made by different firearms during weapon operations. Comparison between different ways of analysis of acoustic signals. Application of machine learning for identification of acoustic signals. Successful identification of 80–90 % of acoustic signals from validation group. Abstract: During the operation of firearms different sounds are made. The sound of a shot, the sound of a flying bullet as well as the sound of the bullet's impact have all been investigated. However, less attention has been given to the sound of the mechanical operation of firearms. This research demonstrates that different types of firearms make different acoustic signals. Moreover, signals which are made by the same firearm during different operations are different. The article discusses how these acoustic signals can be analyzed, compared, and identified in a few different ways. As a result, machine learning has been found to be the most promising for this purpose. The research proves that the presented method of analysis of acoustic signals made by firearms can be used in forensic identification. In addition, the type and stages of further investigation are defined.