On The Exploration Of Recoverable Authentication Scheme For Speech Signal Based On Digital Watermark

The development of high-speed networks and digital speech processing technology make the digital signal as main carrier used for acquiring and exchanging information,such as digital speech signal.Digital speech is easy to store,copy and transmission,which is as information carrier widely used in telephone communications and news stories,as well as the use of electronic evidence in medical records and courts.Nowadays,the increasingly rich of the multimedia editing tools provides convenience for people's life,but make it easy to attack and forge the digital speech.The attacked or forged speech signals used in telephone communications,courts and news stories will distort the facts,impair other's benefit and affect the credibility and public recognition of digital speech seriously.In order to improve the credibility and public recognition of digital speech signal,tamper recovery for the attacked speech has become a research focus in the area of multimedia information security.For the proof in courts,widely used in news stories and actual demand of verifying the authenticity and integrity of digital speech signals,the exploration of forensics and tamper recovery for digital speech will be conducted.The details are as follows:1.A digital speech forensics algorithm robust against feature-analyzed substitution attack is proposed,in order to solve the security problem of the content-based digital watermark schemes.We define the signal's energy ratio of speech signal,and propose the embedding method based on the feature.Frame number is as watermark and embedded,which is used to locate the attacked frames.The scheme has the ability of tamper location precisely,and improves the security of watermark system.2.Based on DWT(Discrete Wavelet Transform),we proposed the compression and reconstruction method for digital speech signal,and the recoverable authentication scheme.Frame number and the compressed signal are as watermark and embedded into speech based on block-based method.If watermarked speech is attacked,frame number is used to tamper location,and compressed signal is used to tamper recovery.The scheme proposed has the ability of content authentication and tamper recovery,improve the credibility and public recognition of digital speech signal.