Research On Perceptual Hashing Authentication Algorithm For Multi-format Audio

It is necessary to accomplish the authentication of speech content integrity and authenticity certification because of speech information is vulnerable by malicious tampering and wiretap.However,the existing algorithms can only be used to complete the authentication for single audio format,at the same time,its efficiency is low,its security is poor,and the tamper detection and positioning cannot be achieved well.It is powerless to achieve audio content security certification of multi-format audio with high efficient.As a result,a perceptual hashing certification method is necessary for multi-format audio with high-efficiency and great security.Aiming at the existing problems of audio authentication algorithms,such as problems that audio files are kept in a single format and algorithms are not generic,poor security,low efficiency,and cannot complete small-scale malicious tampering detection and location.The thesis use speech perceptual hashing to accomplish the authentication of content integrity and authenticity,the robustness,discrimination,security,and authentication efficiency,the small-scale malicious tampering detection and location are studied.The main research work is as follows:1.Aiming at the existing problems of speech authentication algorithms,such as problems that audio files are kept in a single format and algorithms are not generic,poor security,low efficiency.An audio perceptual hashing authentication algorithm for multi-format audio based on energy to zero ratio was proposed.Firstly,the audio signal is processed by applying preprocessing,framing,and adding window to calculate the short-time energy and the number of short-time zero-crossing of each audio frame and to obtain the ratio value of them.Secondly,the ratio value of short-time energy and the number of short-time zero-crossing is used as the perceptual feature value of audio signal.Finally,the feature values are used to conduct a hashing structure process to produce the perceptual hashing sequence.In order to improve the security of the algorithm,the perceptual hashing sequence is scrambled by a key which has the same length as the sequence.Experimental results show that the proposed algorithm shows a high efficiency in terms of perceptual robustness,discrimination and time consumption for audio signal with five different formats in uncompressed domain and compressed domain.2.Aiming at the existing problems of audio authentication algorithms,such as problems that audio files are kept in a single format and algorithms are not generic,poor security,low efficiency,and cannot complete small-scale malicious tampering detection and location.A multi-format audio perceptual hashing authentication algorithm based on discrete wavelet transform(DWT)and measurement matrix was proposed in this thesis.Firstly,the speech signal is conducted with DWT by applying preprocessing,and the low-frequency wavelet coefficients are regarded as the perceptual feature value.Then the measurement matrix which controlled by Logistic chaos map is applied to reduce the dimension of feature value.Finally,the feature value is used to generate the perceptual hash sequence by the process of hashing structure.The measurement matrix based on Logistic chaos is designed as the secret key to enhancing the security of the proposed algorithm.Experimental results show that the proposed algorithm demonstrates a high efficiency in terms of perceptual robustness,discrimination,time consumption and security for audio signal with eight different formats,as well as having a high accuracy of tampering detection and localization.