An information hiding algorithm for MP3 audio on VLC domain

Abstract

This work presents an information hiding algorithm for MP3 audio on VLC domain. The proposed algorithm obtains a Huffman table of maximal probability by analyzing the Huffman tables transcendental probability for a larger numbers of MP3 audios. The variable length codes (VLC) in the obtained Huffman table, which meet the equality of length of codeword and sign bit, are divided two opposite codespaces. The VLC belonging to the two codespaces could denote one binary information correspondingly. According to the information taken in VLC, secret information is embedded to variable length codes. When secret information is extracted, it only needs judging the VLC information without original MP3 audio. Experimental results show that the proposed algorithm can obtain higher hidden data capacity with lower computational complexity and take on excellent imperceptibility.