A Steganalysis Method of LSB Matching for MIDI Audio
Cui Wang · Information Security and Communications Privacy · 2010
In the paper,a method of steganalysis is proposed for MIDI audio by LSB Matching based on the rate of smoothness——of which the orginal audio is different from the audio with hidden message. Hidden message would reduce the smoothness of stego MIDI audio. A conversion rate threshold value is picked up by training a set of MIDI audio for distinguishing whether a MIDI audio is Embedded. Experiments show that,when embedded rate is more than 10%,steganalysis in MIDI audio by the proposed method could be of high correct decision rate.