CN-114663270-B - Image visual safety method based on compressed sensing and discrete wavelet transformation

Abstract

An image visual safety method based on compressed sensing and discrete wavelet transformation belongs to the technical field of information safety. The invention designs an image visual safety algorithm based on compressed sensing and discrete wavelet transformation aiming at the problems that after the image is only suitable for compressed sensing and simple encryption, the histogram is not uniform enough, the information entropy is not close to an ideal value, and the visual meaning is not available. According to the method, a one-dimensional chaotic equation is used as a pseudo-random sequence generator, arnold cat mapping is utilized to scramble an original image, a compressed sensing technology is used for compressing the scrambled image, then a shift exclusive OR operation is used for replacing pixel values, so that the safety is improved, and finally an intermediate state ciphertext is embedded, so that a visual safety image is obtained. Through theoretical analysis and experimental simulation, the invention has excellent encryption performance, the embedding method can achieve visual safety, can realize lossless extraction and image reconstruction after embedding, and has wide application value in a real-time transmission network.

Inventors

• TONG XIAOJUN
• PAN TAO
• ZHANG MIAO
• WANG ZHU

Assignees

• 哈尔滨工业大学（威海）

Dates

Publication Date

20260508

Application Date

20220307

Claims (1)

1. 1. The image visual safety method based on compressed sensing and least significant bits is realized by the following seven steps: firstly, selecting an image X with the size of MxN, carrying out hash operation on the image X to obtain a string of 256-bit binary numbers, converting the binary numbers into decimal numbers and carrying out decimal operation to obtain an initial key X 0 of an equation; secondly, inputting an initial key x 0 into a one-dimensional high-randomness chaotic equation, iterating for 4 times X M times N times, discarding the number of the front M times N in order to ensure better randomness of the pseudo-random sequence, equally dividing the pseudo-random sequence with the length of the rear 3 times X times N into three sections to obtain { x n },{y n },{z n }, and then generating a three-section key stream The specific generation method is as follows: thirdly, performing multidimensional discrete wavelet transformation on the image X to obtain wavelet coefficients, setting a threshold value T=0.080, and performing zero setting on the wavelet coefficients smaller than the threshold value T to further sparse so as to obtain a sparse matrix D; Fourthly, arnold transformation is carried out on the sparse matrix D, and the specific method is as follows: the matrix coordinate position is the transformed matrix coordinate position; Fifth, a measurement matrix is generated by using a pseudo random sequence { x n }, a compression ratio is set to be CR, a measurement matrix phi with the size of CR multiplied by M multiplied by N is obtained, then compressed sensing measurement is carried out on a sparse matrix D, and the calculation formula is as follows: Y=ΦD (5) thereby obtaining a scrambled matrix D' of size CR×M×N; sixth, the scrambled matrix D' is converted into a one-dimensional array, and diffusion operation is prepared, wherein the specific operation is as follows: E i ＝D i-1 ＜＜＜mod(x i *10 14 ,4) (7) Wherein i=2,..cr×m×n, P is an intermediate ciphertext image; Seventh, selecting carrier images with the size of M 'x N', and embedding the intermediate ciphertext images into the carrier images by using a method of least significant bits, wherein the specific method comprises the following steps: the carrier image is subjected to discrete wavelet transformation to obtain four high-low frequency components, then selective embedding is carried out according to the average value of the high-frequency components, and the intermediate state ciphertext image is respectively embedded into the high-intermediate frequency component or the intermediate-low frequency component at the position which is larger than or smaller than the average value, so that the visual safety image is finally obtained.

Description

Image visual safety method based on compressed sensing and discrete wavelet transformation Technical Field The invention belongs to the technical field of information security, and particularly relates to an image visual security method based on compressed sensing and discrete wavelet transformation. Background With the development of information technology in the Internet age, people increasingly use pictures to acquire and transmit information, in real-time communication transmission, direct encryption is sometimes carried out for ensuring the privacy of the pictures, others and attackers cannot decipher the pictures, and in order to cover the attention of the attackers, an encrypted picture needs to be embedded into a carrier image through an embedding technology to ensure the visual safety of important information, so that research on an encryption method with high safety and an embedding method with high feasibility is an important research method. The compressed sensing can realize the nyquist sampling rate without losing important information, so that the compressed sensing is widely used in the field of image signal processing, can realize the rapid compression of images and can recover main information of the images to the greatest extent, but is not safe enough when encountering statistical attack, and other encryption technologies are often used before and after compression. In the past, people have generally encrypted text or converted images into one-dimensional data by shifting, value changing, etc., which, however, has resulted in neglecting the two-dimensional nature of the images, and security cannot be ensured at the time of transmission. Therefore, at present, many novel encryption technologies adopt different encryption technologies such as chaotic systems [1-5], DNA encoding technologies [6-8], cellular automata [9,10], matrix encoding technologies [11,12], optics [13,14] and quantum conversion [15,16] to encrypt images into snowflake-shaped ciphertext images, but the technologies cannot achieve visual safety, so that attention of some attackers is drawn. The visual security algorithm is a method combining compression encryption and embedding, and the method is that an image is compressed and encrypted and then embedded into a carrier image, so that an attacker cannot judge whether the image carries secret information, and the transmission security is improved to a greater extent. For example, chai et al [17] uses compressed sensing and zigzag confusion to perform compressed encryption, uses DWT decomposition on a carrier image, and embeds the decomposition coefficient to obtain a visual security image. Hua et al [12] proposes a visual security image encryption scheme of self-adaptive threshold sparsification and PCS, and embeds a secret image into a carrier image by using a matrix coding technology, so that a reconstruction image with higher quality can be recovered, the distortion rate of the carrier image is reduced by using the matrix coding technology, and the security of an encryption algorithm is stronger. The invention researches a compressed sensing and discrete wavelet embedding method, aims to improve the visual quality of reconstruction after image compression, and simultaneously uses an encryption algorithm with higher complexity to ensure the safety of intermediate state ciphertext of an image, and the embedded visual safety image is highly similar to a carrier image in structure to ensure the visual safety of the image. Disclosure of Invention The invention aims to encrypt a large number of images into a snowflake-shaped ciphertext image by using a traditional encryption technology, but the method attracts attention of an attacker when the images are transmitted in a channel in real time and is easy to crack, so that the safety is increased by researching a method with visual significance. The invention discloses an image visual security method based on compressed sensing and discrete wavelet transformation, which comprises collective algorithm content and specific implementation modes. The technical scheme adopted for solving the technical problems is that an image visual safety method based on compressed sensing and discrete wavelet transformation is provided. The scheme mainly comprises two modules, namely compression encryption and discrete wavelet transform embedding. 1. Chaotic system and pseudo-random generator The one-dimensional Logistics chaotic system and the one-dimensional sine mapping are two classical one-dimensional chaotic systems, and a new highly random one-dimensional chaotic system is designed based on the two classical chaotic systems, and the equation is expressed as follows: xn+1＝a sin(π2a2xn(1-xn))(a-sin(π2a2sin(πxn(1-xn)))) (1) Wherein a is a control parameter, when the control parameter is more than 200, the Lyapunov index reaches 24, and the Lyapunov expresses the orbit separation speed, which indicates that the randomness of the equation is st