Research Of Speech Hiding Algorithm Based On Chaotic Encryption And DWT-DCT In The Covert Communication

In the field of information security, information secretly transmitted mainly uses encryption technology to protect. But the conventional packet encryption technology for information process easily attracts the attention of the enemy, so it is possible for an attacker to crack the secret information. With the growing of computer operation ability, deciphering the traditional cryptograph is becoming simpler. Even if not to decipher, we can also destroy or interfere with the communication process directly. Information hiding as one kind of secure communication technology with excellent camouflage, it can hide the existence of secret information to achieve secure communication. Therefore, whether for civilian or for military communications, research on information hiding technology has a profound practical significance.This paper focuses on the study of chaotic encryption, discrete wavelet transform(DWT), discrete cosine transform(DCT) and singular value decomposition(SVD) of speech information hiding algorithm, and test the performances of the algorithms. The main contents of the full paper are as follows:(1) It describes in detail the basic theory of information hiding, the technical performance requirements, the main application areas and current research status. It analyzes the auditory masking effects of human ears, deeply analyzes the theoretical model of audio information hiding. It summarizes the existing audio hiding algorithms and their characteristics, and gives the general evaluation criteria of the hiding algorithm.(2) In this paper, the chaotic sequence is used to encrypt speech signal, and the commonly used chaotic sequences are briefly introduced. The implementation process of chaotic encryption is studied in detail. We mainly analyze the audio information hiding algorithm based on chaotic encryption and DWT which enables the confidentiality of hidden location. Covert experiments and attack tests show that the algorithm has a good perception of transparency and robustness.(3) In this paper, we propose two kinds of speech hiding algorithms. The first kind of speech hiding algorithm is implemented based on SVD and quantization in the DWT and DCT domain. In this algorithm, the secret information is embedded into the DWT sub-band of each speech frame, so as to obtain the maximum singular value of the DCT coefficients. DWT has spatial-frequency localization properties, and DCT has energy compression characteristics, therefore the transparency of the hidden information is improved and the singular value varies little after the attacks from different types of common signals. Another algorithm is proposed based on the first algorithm. Before embedding the secret speech, we first preprocess each speech frame using Logistic chaotic sequence, to improve its privacy and security during the transmission process. Then we make three levels of DWT to the carrier speech signal, segment the obtained low frequency components, and perform DCT on each segment. Eventually the chaotic encrypted secret speech information is embedded into the intermediate frequency coefficients of the DCT domain. According to the perception redundancy of human auditory system, the secret information embedded into the low-frequency components of the mixed domain has a better perceptual transparency. The simulation results show that this algorithm has a good perception of transparency, and it is able to resist the common signal processing attacks, and has high hiding capacity and good security. Two kinds of algorithms do not require the original carrier speech signal when they extract the secret information, and thus realize the blind extraction of secret information.