Analyzing Texture Features of Ballistic Projectile Images for Firearm Identification
Faculty of Computing, Health and Science
School of Computer and Security Science
Ballistic projectile imaging analysis is a challenging task in forensic identification. The fundamental idea behind firearm identification is that the striations and markings left on fired bullets and cartridge cases are distinct. Traditional firearm identification is conducted by comparing the ballistic projectile characteristics with respect to the lands and grooves, barrel deformities, and striations. The efficiency of traditional firearm identification is heavily dependent on the expertise and experience. So, intelligent identification is highly demanded for effective firearm identification. This paper presents a novel measure criterion for identifying firearm by analyzing the texture features of ballistic projectile images. In doing so, we employ the line-scan optical system for digitizing cylindrical ballistics specimens into 2D images. The texture features of 2D images are then quantified using statistical and spectral techniques. Experimental studies demonstrate that the proposed approach is capable of detecting the firearm efficiently and effectively.