Research On Speech Content Authentication Based On Adaptive Word-framin

With the development of multimedia technology,the acquisition,dissemination,modification and other operations of multimedia information have become extremely convenient and common,resulting in the reduction of the cost of tampering multimedia information.Therefore,it is urgent to adopt technical means to authenticate the integrity and security of multimedia information.Digital watermarking is a technology that protect the copyright,integrity and traceability of products,which is achieved through embedding digital signal into another covered digital product by embedding algorithm.It has attracted great attention of researchers and has been used in many multimedia signal processing fields.Digital watermarking technology is an effective means to identify the integrity of speech.By embedding watermark in speech frame,the integrity of speech content can be well guaranteed,and the untampered speech content can be identified.This paper is devoted to the research of content integrity in speech based on digital watermarking.The work is as follows:1.In order to solve the problem of malicious tampering of speech,the speech feature is embedded into the speech as watermark to ensure the integrity of the speech content.Most of algorithms for authenticating speech content integrity through digital watermarking take fixed-length framing to implement watermark embedding,which weakens the audibility of speech.Based on this,this paper proposes a speech content authentication algorithm based on adaptive word framing.The speech frames are found by the endpoint detection method,and the approximation coefficients and detail coefficients are obtained by discrete wavelet transform.Then,multiple features are extracted from the approximate coefficients,and the watermark is generated by feature fusion and difference value hash.The watermark is embedded by modifying the detail coefficients.In the part of speech content authentication,firstly,the speech frame is segmented according to the adaptive word framing,and the watermark information is reconstructed and extracted for each speech frame in turn.The tamper location is performed by comparing the reconstructed watermark and the extracted watermark.Experimental results show that the framing method of the used has strong robustness.At the same time,the watermark is sensitive to malicious tampering and conventional processing,and has good imperceptibility,which can realize the tamper location of a variety of malicious attacks,and the tamper location accuracy is high.2.In order to make full use of speech and recover tampered speech content on the basis of realizing tampered location,this paper proposes a speech content authentication and tamper recovery scheme based on adaptive word framing.The endpoint detection method is used to find the speech frame,and a hybrid compression algorithm is designed to obtain the compressed data of the speech frame.Secondly,the compressed data is embedded into the least significant bits of the speech frame as a watermark.Finally,the integrity of the speech content was verified and the tamper location was realized by watermarking.In addition,in view of the fact that the existing speech content authentication schemes can only achieve tamper localization and cannot recover the tampered content,a tamper recovery algorithm based on hybrid compression algorithm was proposed to recover the tampered speech frame through compressed data.A large number of experiments show that the scheme has good imperceptibility and sensitivity,and can realize tamper location and recovery from a variety of malicious attacks,and the recovery effect is good.