Research On Watermarking Technology Of Digital Speech Authenticaiton

With the development of the Internet and digital technology, the digital multimedia's production, acquisition and transmission become more and more convenient. The widely using of edition software leads to that multimedia data can be easily tempered under low cost, and the authenticity and integrity of the speech files are often doubted. As an effectively complementary method for traditional security scheme, such as encryption, digital watermarking is a new technology for copyright protection and integrity authentication over the open networks.Recently, digital watermarking technology has been gained a great attention and it has become a hot topic in the field of multi-media signal processing. To solve content authentication problem, a new technology—fragile watermarking, has been proved to be a promising solution. This paper mainly focuses on designing speech fragile watermarking algorithm for content authentication, the main work is given as follows:(1) In order to overcome the weakness of tamper localization which needs the original watermark in fragile watermarking, a fragile watermarking algorithm for speech content authentication is proposed. By quantization, binary watermark image is repeatedly embedded into the wavelet coefficients of the speech signal. This algorithm can locate tampered region without the original watermark. The tamper localization can be realized by extracting two watermarks versions, and calculating the similarity. Experimental results show that the fragile watermarking scheme has the advantages such as low computational complexity, high accuracy of tamper detection and location against cuttings inserting and so on; In addition, it is sensitive against adding noise, re-quantization, low-pass filtering and MP3compression, etc.(2) Aiming at the security problem of using binary image as fragile watermark, a novel fragile watermarking algorithm for speech content authentication is proposed and the fragile watermark is generated from speech itself. Firstly, the generated fragile watermark is embedded into the hybrid domain constituted by discrete wavelet transform and discrete cosine transform for every speech frame in the first speech part. Then, based on the stable characteristics of average value,16bits synchronization code was embedded in the second section for each speech-frame. Experimental results show that the fragile watermarking scheme is not only inaudible and sensitive against additive noise, re-quantization, etc, but also fragile to malicious modification. Especially, it can detect and localize the tampered regions effectively.