Steganalysis: Detecting Existence of File Types Embedded in Grey Scale Image

• Main Author: Amirulikhsan, Zolkafli
• Format: Thesis
• Language: English; English
• Published: 2008
• Subjects: QA76 Computer software
• Online Access: http://etd.uum.edu.my/962/1/Amirulikhsan_Zolkafli.pdf; http://etd.uum.edu.my/962/2/Amirulikhsan_Zolkafli.pdf; http://etd.uum.edu.my/962/

Steganography is an art and science of hiding secret messages into other digital mediums in such a way that no one apart of the intended recipient knows the existence of the message hidden. Recently, steganography have been linked with terrorism activities prior to the September 11th 2001 tragedy. There have been claims that a terrorist group known as Al-Qaeda has been using steganography as a way of secretive communication in spreading their ideologies and attack plans by hiding malicious materials in digital images via the internet. To intercept possible current and future threats, computer forensic examiners must try to identify and intercept any possible secret communication. In doing so, not only the capability to detect the existence of hidden messages is vital but efforts to determine the nature of the hidden message and extract it are also needed so that any malicious plans in the future could be predicted and intercepted earlier. Active steganalysis is presented as ways to expose the possibility of detecting existence of hidden messages, hence defeating steganography. Estimating some parameters of embedding algorithm is the approach that the study used since secret message is one of its essential parameters. Experimental research method was used to detect hidden messages with various file formats in grayscale digital images by conducting two experiments. The various file formats will act as the different possible types of hidden messages that could have been used by terrorists. Here we show that digital images embedded with different file format leaves unique statistical evidence that could be used for detection and estimating the nature of the hidden message with the aid of image processing (image subtraction) and statistical analysis (Paired T and ANOVA).