A Semi-Fragile Digital Audio Watermarking Scheme for MP3-Encoded Signals Using Huffman Data

摘要

Dynamic development in the field of information technologies, storage and sharing a large amount of data via Internet, raises the following question: How to ensure and protect a multimedia document against illegal distribution, reproduction and copying? One of the proposed solutions is digital watermarking which is an approach that embeds imperceptible information into digital data. Although most of the known algorithms operate in the uncompressed or linear domain, few are able to embed watermarks in the compressed domain. This paper describes an audio watermarking scheme that operates directly in the compressed domain. It inserts watermark in MP3 bit streams. Our scheme uses Huffman data mainly big values region and recompression calibration to hide secret information. Furthermore, it allows blind retrieval of an embedded watermark which does not need the original audio. Experimental results prove the imperceptibility of the proposed method and the robustness against several attacks.