|
Activity Number:
|
342
|
|
Type:
|
Topic Contributed
|
|
Date/Time:
|
Tuesday, August 2, 2016 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Defense and National Security
|
|
Abstract #319886
|
|
|
Title:
|
Quantifying Quality and Uncertainty in Forensic Pattern Matching
|
|
Author(s):
|
Lucas Mentch* and Duy Thai
|
|
Companies:
|
North Carolina State University and Statistical and Applied Mathematical Sciences Institute
|
|
Keywords:
|
Image Decomposition ;
Latent Evidence ;
Bootstrapping ;
Quality Metric ;
Nonparametric Testing
|
|
Abstract:
|
A 2009 report by the National Research Council (NRC) shed light on some major weaknesses in forensic science. Outside of DNA matching, nearly every other forensic technique was shown to have potentially serious flaws in both the analysis and reporting of the evidence. This work presents automated methods for assessing quality and evaluating uncertainty associated with latent fingerprints. We begin by describing an image decomposition technique that can be used for defining and extracting regions of poor contrast from raw images. This decomposition is then used to develop standardized quality metrics in an automated fashion that can be used to examine the relationship between the quality scores and the variability in potential matches returned from a database of noise-free images. The methodology developed extends naturally to all forms of latent pattern evidence.
|
Authors who are presenting talks have a * after their name.
