Paper Info
Title
Improved Image Analysis Methodology for Detecting Changes in Evidence Positioning at Crime Scenes
Abstract
This paper proposed an improved methodology to assist forensic investigators in detecting positional change of objects due to crime scene contamination. Either intentionally or by accident, crime scene contamination can occur during the investigation and documentation process. This new proposed methodology utilises an ASIFT-based feature detection algorithm that compares pre- and post-contaminated images of the same scene, taken from different viewpoints. The contention is that the ASIFT registration technique is better suited to real world crime scene photography, being more robust to affine distortion that occurs when capturing images from different viewpoints. The proposed methodology was tested with both the SIFT and ASIFT registration techniques to show that (1) it could identify missing, planted and displaced objects using both SIFT and ASIFT and (2) ASIFT is superior to SIFT in terms of error in displacement estimation, especially for larger viewpoint discrepancies between the pre- and post-contamination images. This supports the contention that our proposed methodology in combination with ASIFT is better suited to handle real world crime scene photography.
Year
DOI
Venue
2019
10.1109/DICTA47822.2019.8945934
2019 Digital Image Computing: Techniques and Applications (DICTA)
Keywords
Field
DocType
Image registration,change detection,crime scene image analysis
Affine transformation,Scale-invariant feature transform,Computer vision,Crime scene,Change detection,Pattern recognition,Viewpoints,Computer science,Photography,Artificial intelligence,Distortion,Image registration
Conference
ISBN
Citations 
PageRank 
978-1-7281-3858-9
0
0.34
References 
Authors
5
3
Name
Order
Citations
PageRank
Mark Petty100.34
Shyh Wei Teng215121.02
M. Manzur Murshed300.34