IMPROVEMENT OF STEGOANALYSIS METHODS FOR DETERMINING TRAINING IN AUDIO FILES

Abstract

The paper considers the improved methodology and the software developed for its implementation, which makes it possible to detect attachments, made with the help of the LSB algorithm, in some types of audio files of the WAVE format of any bit rate and sampling frequency. It also provides us with the possibility to determine the presence of stego attachments in certain types of audio files. The testing procedure is based on methods of determining stego attachments using the frequency method and compression method. The suggested methods of searching for attachments in file service areas and developed software in Embarcadero Delphi XE5 environment effectively reveal the fact of the implementation of the message created by the majority of existing programs. The software was also tested on a large number of files without attachments. As a result a large number of false positives were revealed. The reason for this was damaged structure of MP3 files (corrupted frames were taken for possible stego attachments).