Weapons and Related Object Classification in Digital Forensic Using Machine Learning

Weapons and Related Object Classification in Digital Forensic Using Machine Learning

연구 분야: Analysis

학회: 2023 14th International Conference on Computing Communication and Networking Technologies (ICCCNT)

The use of machine learning algorithms for the bracket of munitions and affiliated objects has come increasingly important in the field of digital forensic. This paper presents a comprehensive review of recent studies on the operation of machine learning ways (Convolutional Neural Network) to identify munitions and affiliated objects from multiple sources, including images of different of guns, cutter, vest etc. The review highlights the strengths and limitations of machine learning approaches, similar as unsupervised and pre-trained model learning. The performance of these models in terms of accuracy, sensitivity, and specificity is also estimated. The paper discusses several issues that need to be addressed in the operation of machine learning models to forensic examinations, similar as the need for large annotated datasets, robust point birth styles, and interpretability of the models. The paper concludes with a discussion of unborn exploration directions and the implicit impact of machine learning on the forensic analysis of munitions and affiliated objects.