Research On Retrieval And Authentication Method For Encrypted Speech Based On Audio Fingerprint

With the rapid development of cloud storage and Internet technology,as well as the explosive growth of the number of Internet audio,high-speed retrieval in audio big data has become an urgent problem to be solved.In addition,the audio data stored in the cloud saves local space for users,but it also brings problems such as finding difficulties,privacy leakage,and data insecurity.Therefore,how to improve the retrieval efficiency of speech files,ensure the data security of speech files,and ensure the integrity and authentication accuracy of encrypted speech has important research significance.Aiming at the problems of efficient retrieval and content authentication of encrypted speech,this thesis uses audio fingerprint,feature dimension reduction,speech encryption,robust watermarking and fragile watermarking technologies to study key technologies such as encrypted speech retrieval and content authentication.The main research work is as follows:1.In order to solve the problems of the existing audio fingerprint method when extracting audio fingerprints from long speech segments,such as too large fingerprint dimension,poor robustness,and low retrieval accuracy and efficiency,a robust audio fingerprint retrieval method based on feature dimension reduction and feature combination is proposed.Firstly,the Mel-frequency cepstral coefficient(MFCC)and linear prediction cepstrum coefficient(LPCC)of the original speech are extracted respectively,and the MFCC feature matrix and LPCC feature matrix are combined.Secondly,the feature dimension reduction method based on information entropy is used for column dimension reduction,and the feature matrix after dimension reduction is used for row dimension reduction based on energy feature dimension reduction method.Finally,the audio fingerprint is constructed by using the feature combination matrix after dimension reduction.In the retrieval,the normalized Hamming distance algorithm is used for matching retrieval.Experiment results show that the proposed method has smaller audio fingerprint dimension and better robustness for long speech segments,and has higher retrieval efficiency while maintaining a higher recall rate and precision rate.2.In order to improve the retrieval efficiency and accuracy of encrypted speech and verify the legality of the retrieved speech,an encrypted speech retrieval method based on audio fingerprints is proposed.Firstly,the audio fingerprint is embedded into the original speech as digital watermarking by using the robust audio watermarking algorithm,then the original speech with watermark is encrypted by scrambling encryption algorithm based on Henon chaotic map,the frame number of the encrypted speech is embedded into the encrypted speech by using the fragile watermarking algorithm,and the encrypted speech database is constructed.Finally,the one-to-one mapping relationship between each audio fingerprint and the corresponding encrypted speech is established to upload the cloud.The experimental results show that the encrypted speech retrieval algorithm has excellent retrieval performance and higher retrieval efficiency,and the robust watermarking algorithm has high embedding capacity,transparency and robustness,which can well meet the needs of information annotation.3.In order to authentication the integrity of the encrypted speech with retrieval feedback,and solve the problem that most of the existing speech authentication algorithms can only realize tamper location of frames and need to the additional recovery signals during tamper recovery,this thesis proposes a encrypted speech authentication and recovery scheme based on fragile watermarking.In this scheme,fragile watermarking information is constructed from the frame number of encrypted speech,and compared with the fragile watermarking information extracted from the speech to be authenticated to achieve tamper detection,and the tampered sample points are determined by locating the detected speech step by step.Finally,the linear interpolation method based on residual error is used to restore the tamper of unauthenticated speech.The experimental results show that this scheme can realize the tamper detection and location of sampling points,and the tamper content after recovery has good auditory quality.