Audio frequency camouflage and recovery method that combines secret key with random unit orthogonal basis

Abstract

The invention discloses an audio frequency camouflage and recovery method that combines a secret key with random unit orthogonal basis; the method comprises the following steps: firstly dividing a secret audio frequency and a public audio frequency into small segment sequences of the same quantity, and using the secret key to build the random unit orthogonal basis; secondarily, solving the projection of the secret audio frequency small segment sequence on the random unit orthogonal basis so as to fully, effectively and linearly express the secret audio frequency small segment sequence, selecting first k projection factors with large amplitude and energy and recording index positions, and thus forming an index sequence; finely adjusting the public audio frequency and using a high capacity built-in secret writing method to embed the selected projection factors and a line sequence into the corresponding public audio frequency small segment sequence, thus forming a channel public transmission audio frequency; using transform parameters extracted from the channel public transmission audio frequency, and combining with the secret key so as to reconstruct the secret audio frequency. Compared with the prior art, the method can reconstruct the secret audio frequency with different precisions, and strictly depends on the user secret key, thus providing a higher safety.

Description

technical field [0001] The invention belongs to the intersecting field of information security and digital audio signal processing, relates to an audio camouflage recovery method, in particular to an audio camouflage and recovery method combining a secret key and a random unit orthogonal basis. Background technique [0002] In recent years, with the development of computer and network technology, more and more digital images and audios are transmitted in public channels. Unrestricted access and use of secret images and audios not only involves personal privacy, media credibility, and government integrity, but also leads to social unrest and military conflicts. [0003] How to effectively protect the secret image and audio transmitted in the public channel has become a hot research topic at present. Aiming at this problem, people have proposed a variety of image and audio protection methods, such as digital image and audio encryption technology and digital image and audio ca...

AI Technical Summary

Problems solved by technology

[0010] But whether it is ① standard Tangram algorithm, or ② block average 5-tuple matching Tangram method and the previously given ③ improved Tangram algorithm and 2-dimensional double-scale rectangular mapping digital audio camouflage method, ④ digital audio camouflage based on block sequence and reconstruction method, ⑤ digital audio camouflage and reconstruction method based on segmented sequence, and ⑥ audio information hiding method combined with dictionary and sorting linear fitting. The transformation models based on these methods are all affine transformation models. For image , the affine transformation model only has the mean value block and the difference block, corresponding to the audio, the affine transformation model only has the mean value vector and the difference vector, and the mean value block and the difference block as well as the mean value vector and the difference vector do not satisfy the basic orthogonal relationship, resulting in affine transformation The transformation accuracy of the model is generally low, and it cannot effectively guarantee the fitting accuracy of the secret image and audio to the public image and audio, so that effective channel deception cannot be carried out, and at the same time, it cannot effectively guarantee the fitting accuracy of the public image and audio to the secret image and audio. Therefore, it is impossible to guarantee the accurate reconstruction of secret image audio and obtain high-precision secret image audio

On the other hand, if the difference block and the mean value block or the mean value vector and the difference vector of the affine model tend to be consistent, it corresponds to a constant value block or a constant value sequence processing. At this time, random disturbance must be added to improve the secret image audio and Publicly discloses the matching performance of small blocks and segments of image and audio division, otherwise it will not be restored, thus further limiting the transformation accuracy of the radiation transformation model

For (7) the Tangram algorithm based on triangulation can only approximately reconstruct the triangulation area of ​​the secret image, and its practical application value is still small

Images