About CS&CE
Prospective Students
Research
Staff
Dept Comp Sci & Comp Eng
Contact Details
La Trobe University
Victoria 3086
AUSTRALIA
Tel: +61 3 9479 1107
Fax: +61 3 9479 3060
Email: info
@cs.latrobe.edu.au
|
 |
Grants - Detail
Department of Computer Science & Computer Engineering
| Forensic Image Mining: A case for Mining Shoe Marks in Crime Scenes |
The aim of this project is to apply data mining techniques to shoe mark images to extract hidden pattern and knowledge in order to help investigators understand past and new crimes. The Victorian Police Forensic Services Department is supporting this project by contributing their knowledge on crimes, as well as providing us with shoe mark image databases for analysis.
In crime forensic, in order to build a case for the crime, it is very important and necessary to determine and prove that a particular person or persons may or may not have been present at the crime scene. For this reason, crime scene investigators collect and analyze crime scene evidences through scientific and advance technology, such as camera, chemical solutions, etc. There are many different types of physical evidence accumulated in the forensic database like fingerprints, tool marks, footwear evidence, tire marks, blood stains and so on. For this project, we will be focusing on the footwear evidence database, which contains thousands images of shoe marks in the crime scenes. Discovering patterns from the massive amount of shoe mark image data is critical in solving new crime cases, as well as past-unsolved cases, through the discovery of new knowledge and patterns.
Discovering patterns of shoe mark images is an important part of solving criminal cases, as criminals become smarter by starting to wear gloves and masks to avoid leaving traces. However, most of the criminals normally do not attempt conceal their footwear thus leaving some footwear evidence at the crime scenes. Although the shoeprints are sometimes overlooked or destroyed due to the improperly secured and disorganized crime scenes, this type of evidence can be greatly important to crime scene investigators in proving or disproving the presence of a suspect.
The objectives of this proposed projects are as follows:
To perform image cleaning and transformation,
To apply data mining techniques, including clustering, classification, and association rules to the shoe mark image collections, and
To develop a holistic and hybrid technique to incorporate the abovementioned data mining techniques applied to the shoe mark image collections.
This research is significant, since it addresses practical and important problems of forensic image analysis, through the use of advanced techniques in computer science (data mining and image processing). This project combines three fields, namely data mining, image processing, and forensic science, demonstrating the use of computer technology and techniques (i.e. data mining and image processing) for the advancement of a different field (i.e. forensic science); which can be a showcase for an E-Research. Firstly, CI expertise in data mining compliments, secondly, CI experience in forensic projects.
The expected outcome of this research will be the extraction of patterns and knowledge found in the shoe mark image collections, for which this knowledge may help investigators uncover new knowledge about the crimes.
|
|
|
