Related Experiment Video
Updated: Oct 26, 2025

07:13
Author Spotlight: An Efficient and Robust Software for Automated Fusion of Multiple Preclinical Imaging Modalities
Published on: October 27, 2023
1.4K
Statistical models for firearm and tool mark image comparisons based on the congruent matching cells (CMC) method
1National Institute of Standards and Technology, 100 Bureau Dr, Gaithersburg, MD 20899, USA.
Forensic Science International
|July 27, 2021
Summary
This study introduces a generalized frequency function to accurately model firearm evidence identification using the congruent matching cells (CMC) method. The new approach addresses varying numbers of cell pairs in image comparisons for more reliable firearm mark analysis.
Area of Science:
- Forensic Science
- Firearm Evidence Identification
- Ballistics
Background:
- The congruent matching cells (CMC) method is used in firearm evidence identification to compare topography images of breech face impressions.
- CMC analysis quantifies similarity by determining correlated cell pairs and identifying congruent match cells (CMC pairs).
- Accurate estimation of error rates and likelihood ratios (LR) requires appropriate probability distributions for CMC results.
Purpose of the Study:
- To discuss statistical models for CMC measurements, including binomial and related distributions.
- To address the limitation of previous studies assuming a constant number of Bernoulli trials (N) in CMC analysis.
- To introduce a generalized frequency function for CMC values and provide its limiting distribution, accommodating variable N.
Main Methods:
- Discussion of four statistical models for CMC measurements: binomial and three binomial-related distributions.
- Introduction of a generalized frequency function to model CMC values when the number of cell pairs (N) varies.
- Application of nonlinear regression models to estimate parameters using actual CMC values from fired cartridge cases.
Main Results:
- The generalized frequency function effectively depicts the behavior of CMC values, especially when N varies between image pairs.
- The limiting distribution of the generalized frequency function is provided for theoretical analysis.
- Nonlinear regression models successfully estimated parameters using real-world CMC data from firearm evidence.
Conclusions:
- The proposed generalized frequency function offers a more appropriate statistical framework for CMC measurements in firearm identification.
- This methodology improves the reliability of quantifying topography similarity and estimating error rates in forensic firearm analysis.
- The study provides a robust statistical approach for analyzing CMC data from fired cartridge cases.

