Genetic based machine learning system for mp3 steganalysis

مقدم أطروحة جامعية

al-Buwainah, Muhammad Jazi

مشرف أطروحة جامعية

al-Shalabi, Riyad

أعضاء اللجنة

Kanan, Ghassan Jaddu
Uways, Suhayl

الجامعة

الأكاديمية العربية للعلوم المالية و المصرفية

الكلية

كلية نظم و تكنولوجيا المعلومات

القسم الأكاديمي

قسم نظم المعلومات الحاسوبية

دولة الجامعة

الأردن

الدرجة العلمية

دكتوراه

تاريخ الدرجة العلمية

2010

الملخص الإنجليزي

Moving Picture Expert Group-3 MP3 steganography is the art of hiding data into digital MP3 files.

This science represents a threat to the safeguarding of sensitive information and the gathering of intelligence.

Steganalysis is the art of detecting this hidden information.

It is an inherently difficult problem and requires a thorough investigation.

A strong framework for analysis is required for steganalyst and the steganographer to improve its task in the hiding process inside of steganographer and detection process inside of steganalyst.

In this dissertation, we lie down of foundation of a passive Genetic Base Machine Learning GBML framework for analysis of steganography and steganalysis and use this analysis to create practical solutions to the problems of detecting and evading detection.

GBML previously employed in disciplines such as networking, clustering, and classification to provide a natural framework for studying different type of problems.

With this framework, we make statements on the detectability of modern MP3 steganography schemes, develop tools for steganalysis in a practical scenario, and design and analyze a mean of escaping optimal detection.

We develop our detection framework and apply it to detect the steganalysis of MP3StegC, MP3StegZ, MP3StegInfo and Stegangraphy1.8 hiding schemes.

Experiments over a diverse database of MP3 files show our steganalysis to be effective and competitive with the state-of-the-art.

Though our approach has gained satisfying results in the study of steganography, we acknowledge there are problems yet to be solved in order to improve the detection rate.

Results show the Degree Of Confidence DOC of the system on the average is 0.84 for clear files and 0.88 for stego files, while Measure Of Usefulness MOU on the average for MP3StegC, MP3StegInfo, MP3StegZ, and Steganography 1.8 are 0.64, 0.74, 0.73, and 0.90 respectively.

التخصصات الرئيسية

تكنولوجيا المعلومات وعلم الحاسوب

الموضوعات

• الكتابة
• التشفير
• الرمزية
• البيانات

عدد الصفحات

121

قائمة المحتويات

• APA
• MLA
• AMA

لغة النص

الإنجليزية

نوع البيانات

رسائل جامعية

رقم السجل

BIM-307532