A robust firearm identification algorithm of forensic ballistics specimens

• Published in: Journal of Physics: Conference Series
• Main Author: Chuan Z.L.; Jemain A.A.; Liong C.-Y.; Ghani N.A.M.; Tan L.K.
• Format: Conference paper
• Language: English
• Published: Institute of Physics Publishing 2017
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85030705561&doi=10.1088%2f1742-6596%2f890%2f1%2f012126&partnerID=40&md5=dfdd33901005dc0514414554ec2c2c2f

There are several inherent difficulties in the existing firearm identification algorithms, include requiring the physical interpretation and time consuming. Therefore, the aim of this study is to propose a robust algorithm for a firearm identification based on extracting a set of informative features from the segmented region of interest (ROI) using the simulated noisy center-firing pin impression images. The proposed algorithm comprises Laplacian sharpening filter, clustering-based threshold selection, unweighted least square estimator, and segment a square ROI from the noisy images. A total of 250 simulated noisy images collected from five different pistols of the same make, model and caliber are used to evaluate the robustness of the proposed algorithm. This study found that the proposed algorithm is able to perform the identical task on the noisy images with noise levels as high as 70%, while maintaining a firearm identification accuracy rate of over 90%. © Published under licence by IOP Publishing Ltd.