Gunshots detection, identification, and classification: Applications to forensic science

ce proliferation of audio sensors in surveillance, smartphones, and numerous devices has made gunshots-based event detection and forensic analysis critical for prompt police action and crime scene reconstruction. This paper initiates an analysis of the acoustic characteristics of gunshots and the variables affecting them, assessing their applicability and limitations in forensic science. It follows with a comprehensive review of existing literature on gunshots detection, identification, and classification technologies, detailing the critical components of machine learning applications, including dataset construction, feature extraction, and classifier selection. Despite the challenges in comparing diverse algorithms due to differences in data and evaluation criteria, the adoption of deep learning-driven neural networks is poised to become a dominant trend. This study aims to chart new frontiers in security systems and forensic analysis.