Tong Mingsi, Song John, Chu Wei
School of Mechatronics Engineering, Harbin Institute of Technology, Harbin 150001, China; National Institute of Standards and Technology, Gaithersburg, MD 20899 USA.
National Institute of Standards and Technology, Gaithersburg, MD 20899 USA.
J Res Natl Inst Stand Technol. 2015 Apr 29;120:102-12. doi: 10.6028/jres.120.008. eCollection 2015.
The Congruent Matching Cells (CMC) method was invented at the National Institute of Standards and Technology (NIST) for firearm evidence identifications. The CMC method divides the measured image of a surface area, such as a breech face impression from a fired cartridge case, into small correlation cells and uses four identification parameters to identify correlated cell pairs originating from the same firearm. The CMC method was validated by identification tests using both 3D topography images and optical images captured from breech face impressions of 40 cartridge cases fired from a pistol with 10 consecutively manufactured slides. In this paper, we discuss the processing of the cell correlations and propose an improved algorithm of the CMC method which takes advantage of the cell correlations at a common initial phase angle and combines the forward and backward correlations to improve the identification capability. The improved algorithm is tested by 780 pairwise correlations using the same optical images and 3D topography images as the initial validation.
全等匹配单元(CMC)方法是美国国家标准与技术研究院(NIST)发明的用于枪支证据鉴定的方法。CMC方法将诸如已发射弹壳的后膛面印记等表面积的测量图像划分为小的相关单元,并使用四个识别参数来识别源自同一支枪支的相关单元对。通过使用从10个连续制造的滑块的手枪发射的40个弹壳的后膛面印记捕获的3D地形图像和光学图像进行识别测试,对CMC方法进行了验证。在本文中,我们讨论了单元相关性的处理,并提出了一种改进的CMC方法算法,该算法利用了在共同初始相位角处的单元相关性,并结合了正向和反向相关性以提高识别能力。使用与初始验证相同的光学图像和3D地形图像,通过780对相关性对改进算法进行了测试。