Forensic garment analysis has traditionally been a qualitative task. However, as the "CSI effect" grows and expectations for quantitative methods increase, it becomes more challenging for forensic examiners to rely on "expert opinion" without statistical backing. A quantitative approach is needed when matching seized garments with garments captured in surveillance images and videos.
Quantifying Garment Analysis
To address this problem, Quantum Signal researched and developed a clothing pattern analysis system that uses advanced algorithms to mathematically "fingerprint" a given patterned garment and compare this fingerprint to surveillance images. The system enables investigators to calculate the likelihood that a seized patterned garment is the same as a garment in the surveillance image. This compliments expert examination by providing statistical figures to convey how useful the comparison is.
Quantum Signal first applied this method to U.S. military camouflage uniforms in 2007 and later expanded the effort to the broader class of patterned garments in 2010. For each project, QS created a software application to assist forensic examiners in calculating the statistical likelihood of a match.
These applications are a significant advancement in forensic science. They are the first tools to enable examiners to quantify strength of evidence in a meaningful and rigorous fashion.
Technology in Action:
Military Uniform Uniqueness Statistical Evaluator (MUUSE)
MUUSE is a forensic software system for camouflage clothing pattern analysis that statistically characterizes the uniqueness of individual uniforms and provides investigators with a wealth of camouflage uniform-related forensic information. This software provides forensic experts with a way to easily and accurately perform camouflage uniform matching and analysis and generate reports suitable for use in investigations and legal proceedings. For more information about MUUSE, check out the MUUSE brochure (PDF).
Garment Match LR Calculator
The Garment Match LR Calculator is a new application for patterned garment analysis that allows investigators to determine the numeric likelihood ratio (LR) of a match between a seized garment and corresponding still images of a garment from a surveillance camera. This tool enables examiners to easily and accurately perform a statistical analysis and generate a detailed report suitable for use in investigations.