Using artificial neural networks for forensic file type identification

Get BibTex-formatted data

Download

PDF

Author

Ryan M. Harris

Tech report number

CERIAS TR 2007-19

Entry type

mastersthesis

Abstract

Current forensic software relies upon accurate identification of file types in order to determine which files contain potential evidence. However, current type recognition mechanisms are susceptible to simple attacks that enable a criminal to confuse the detection algorithm. This study investigated whether artificial neural networks were superior to existing mechanisms at responding to modern evidence tampering techniques and concluded that the tested neural networks were not better than the existing methods. However, the study yielded avenues for future investigation.

Download

PDF

Date

2007 – 05 – 11

URL

http://web.ics.purdue.edu/~rmharris/Thesis.pdf

Institution

CERIAS

Key alpha

Harris

Publisher

Purdue University

School

Purdue University

Publication Date

2007-05-11

Subject

Investigates using artificial neural networks to identify a file's content type

BibTex-formatted data

To refer to this entry, you may select and copy the text below and paste it into your BibTex document. Note that the text may not contain all macros that BibTex supports.

@Mastersthesis{ Harris,
	title = "Using artificial neural networks for forensic file type identification",
	author = "Ryan M. Harris",
	year = "2007",
	month = "05",
	institution = "CERIAS",
	publisher = "Purdue University",
	school = "Purdue University",
	day = "11",
	abstract = "Current forensic software relies upon accurate identification of file types in order to determine which files contain potential evidence. However, current type recognition mechanisms are susceptible to simple attacks that enable a criminal to confuse the detection algorithm. This study investigated whether artificial neural networks were superior to existing mechanisms at responding to modern evidence tampering techniques and concluded that the tested neural networks were not better than the existing methods.  However, the study yielded avenues for future investigation.",
	subject = "Investigates using artificial neural networks to identify a file's content type",
	url = "http://web.ics.purdue.edu/~rmharris/Thesis.pdf",
}