Multi-image encryption method based on dynamic stereo S-box and lifting wavelet transform

Abstract

A multi-image encryption method based on dynamic stereo S-box and lifting wavelet belongs to the field of information encryption. At present, the network transmission and storage of digital images are becoming more and more frequent. In order to protect the content of network interactive images, the present invention establishes a dynamic stereo S-box and combines wavelet transform to propose a multi-image encryption method based on dynamic stereo S-box and lifting wavelet transform. First, k The original images are combined into an image cube, and then the lifting wavelet transform is performed; secondly, chaos is used to scramble the low-frequency part in the frequency domain; thirdly, the dynamic stereo S-box is used to replace the pixel values ​​of the image cube after the lifting wavelet inverse transform; finally, the image cube is decomposed to produce k The experimental results show that this method can encrypt multiple interactive images at the same time, with good encryption effect and strong security.

Description

Technical Field

[0001] The method relates to an information encryption technology, in particular to an image encryption method. Background Art

[0002] In the interaction of network information, images are simple, intuitive, and easy to communicate across languages ​​in different regions and ages. They are the main carrier of information dissemination today and have been widely used in various fields. However, image information security issues have also arisen. Therefore, it is particularly important to study efficient and secure image encryption methods.

[0003] In order to improve the security and efficiency of image encryption, a dynamic stereoscopic S-box is established, and a multi-image encryption method based on dynamic stereoscopic S-box and lifting wavelet transform is proposed in combination with lifting wavelet transform. This method uses the nonlinearity of S-box and the good randomness and complexity of chaos to achieve dual encryption in frequency domain and space domain, effectively protecting the security of network transmission and storage of image content. Summary of the invention

[0004] The purpose of the present invention is to solve the problems of low efficiency, weak security or limited capacity of existing image encryption, and propose a multi-image encryption method based on dynamic stereo S-box and lifting wavelet transform.

[0005] The multi-image encryption method based on dynamic stereo S-box and lifting wavelet transform is characterized by comprising the following steps:

[0006] Step 1: Cube the image: Let k The size is m × n The original image is I 1 , I 2 , …, I k , combining them in a specific order into a m × n × k Image Cube C ;

[0007] Step 2: Lifting wavelet transform: C Perform lifting wavelet transform and get C The frequency domain matrix C 1. Order C The low frequency part of 1 is V , whose size is m / 2× n / 2× k ;

[0008] Step 3: Generate a chaotic sequence: Randomly select the initial value of the Logistic map x 0 and parameters μ 1. Follow formula (1) and iterate 1000+ times mnk / 4 times, discarding the first 1000 values, we get a length of mnk / 4 Chaos Sequence X ={ x i}; Randomly select the initial value of the Logistic map y 0 and parameters μ 2. According to formula (1), iterate 1000+256 times and discard the first 1000 values ​​to obtain a chaotic sequence with a length of 256. Y ={ y j};

[0009] X t+1 = x t × μ ×(1- x t ), (1)

[0010] in, μ ∈(3.5699, 4) and x t ∈(0, 1);

[0011] Step 4: Low frequency scrambling: Use formula (2) to X Sort in ascending order,

[0012] [ X 2, X 1]= sort ( X ), (2)

[0013] in, sort (·) is the ascending sorting function, X 1 is the index value sequence, X 2 is the sorted sequence; using X 1 and formula (3) V The elements in the array are shuffled.

[0014] v 1( i )= V [ x 1( i )], i =1, 2, …, mnk / 4, (3)

[0015] in,x 1( i )∈ X 1 and V 1={ v 1( i )} is the low-frequency part after scrambling; V 1 and C The high frequency part of 1 is combined to obtain the frequency domain matrix C 2;

[0016] Step 5: Inverse lifting wavelet transform: C 2 Perform the lifting wavelet inverse transform to obtain the scrambled time domain cube C 3;

[0017] Step 6: Generate dynamic stereo S-box: Use formula (4) to Y Sort in ascending order,

[0018] [ Y 2, Y 1]= sort ( Y ), (4)

[0019] in, Y 1 is the index value sequence, Y 2 is the sorted sequence; using formula (5) Y 1 is converted into a 8×8×4 three-dimensional matrix, i.e. a dynamic stereo S-box;

[0020] S = reshape ( Y 1,8, 8, 4), (5)

[0021] in, reshape (·) is the reshaping function, S It is an 8×8×4 S-box;

[0022] Step 7: Cube Diffusion: C Each pixel value in 3 d Represented in 8-bit binary d=d 7 d 6 d 5 d 4 d 3 d 2 d 1 d 0 , d l∈{0,1}, l= 0, 1, …, 7, let x=d 7 d 6 d 5 , y=d 4 d 3 d 2 , z=d 1 d 0 , by using x , y , z The decimal value of the query element in the S box is used to replace the value of the element C Pixel value in 3 d , after replacing all pixels, we can get the encrypted image cube C 4;

[0023] Step 8: Cube Decomposition: C 4 decomposes into k The size is m × n Encrypted image of J 1 , J 2 , …, J k .

[0024] Beneficial effects: In view of the problems of low encryption efficiency, weak security or limited capacity in existing image encryption methods, this paper proposes a multi-image encryption method based on dynamic stereo S-box and lifting wavelet. The main contributions are: (1) a dynamic stereo S-box is established to improve randomness and security; (2) combined with lifting wavelet transform, dual encryption protection in frequency domain and space domain is realized; (3) this method has the characteristics of high efficiency, security and good encryption effect, which can effectively protect the security of network transmission and storage of image content. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 :Multi-image encryption flow chart based on dynamic stereo S-box and lifting wavelet transform;

[0026] Figure 2 : original image set;

[0027] Figure 3 : Schematic diagram of dynamic stereo S-box;

[0028] Figure 4 : Encrypted image set. DETAILED DESCRIPTION

[0029] The implementation process of the present invention is further described in detail below with reference to specific drawings and examples.

[0030] Figure 1 It is the encryption flow chart of this method.

[0031] The programming software used is Matlab R2020a, and the Figure 2 The 10 images of size 512×512 are shown as the original image set.

[0032] Step 1: Image cubing: Let 10 original images of size 512×512 be I 1 , I 2 , …, I 10 , combining them in a specific order into an image cube of size 512×512×10 C .

[0033] Step 2: Lifting wavelet transform: C Perform lifting wavelet transform and get C The frequency domain matrix C 1. Order C The low frequency part of 1 is V , its size is 512 / 2×512 / 2×10=655360.

[0034] Step 3: Generate a chaotic sequence: Randomly select the initial value 0.253 and parameter 3.891 of the Logistic map, iterate 1000+655360 times according to formula (1), discard the first 1000 values, and obtain a chaotic sequence with a length of 655360. X ={ x i}; Randomly select the initial value 0.126 and parameter 3.725 of the Logistic map, iterate 1000+256 times according to formula (1), discard the first 1000 values, and get a chaotic sequence with a length of 256 Y ={ y i}.

[0035] Step 4: Low frequency scrambling: Use formula (2) to X Sort in ascending order, where X 1 is the index value sequence, X 2 is the sorted sequence; using X 1 and formula (3) V The elements in the array are shuffled, where xi ( i )∈ X 1 and V 1={ v i ( i )} is the low-frequency part after scrambling; V 1 and C The high frequency part of 1 is combined to obtain the frequency domain matrix C 2.

[0036] Step 5: Inverse lifting wavelet transform: C 2 Perform the lifting wavelet inverse transform to obtain the scrambled time domain cube C 3.

[0037] Step 6: Generate dynamic stereo S-box: Use formula (4) to Y Sort in ascending order, where Y 1 is the index value sequence, Y 2 is the sorted sequence; using formula (5) Y 1 is converted into a 8×8×4 three-dimensional matrix, namely a dynamic stereo S-box, such as Figure 3 shown.

[0038] Step 7: Cube Diffusion: C Each pixel value in 3 d Represented in 8-bit binary d=d 7 d 6 d 5 d 4 d 3 d 2 d 1 d 0 , d l ∈{0,1}, l= 0, 1, …, 7, let x=d 7 d 6 d 5 , y=d 4 d 3 d 2 , z=d 1 d 0 , by usingx , y , z The decimal value of the query element in the S box is used to replace the value of the element C Pixel value in 3 d , after replacing all pixels, we can get the encrypted image cube C 4.

[0039] Step 8: Cube Decomposition: C 4 is decomposed into 10 encrypted images of size 512×512 J 1 , J 2 , …, J 10 ,like Figure 4 shown.

Claims

1. A multi-image encryption method based on dynamic stereo S-box and lifting wavelet transform, characterized in that: The steps include: Step 1: Cube the image: Let k The size is m × n The original image is I 1 , I 2 , …, I k , combining them in a specific order into a m × n × k Image Cube C ; Step 2: Lifting wavelet transform: C Perform lifting wavelet transform and get C The frequency domain matrix C 1. Order C The low frequency part of 1 is V , whose size is m / 2× n / 2× k ; Step 3: Generate a chaotic sequence: Randomly select the initial value of the Logistic map x 0 and parameters μ 1. Follow formula (1) and iterate 1000+ times mnk / 4 times, discarding the first 1000 values, we get a length of mnk / 4 Chaos Sequence X ={ x i }; Randomly select the initial value of the Logistic map y 0 and parameters μ 2. According to formula (1), iterate 1000+256 times and discard the first 1000 values ​​to obtain a chaotic sequence with a length of 256. Y ={ y j }; X t+1 = x t × μ ×(1- x t )， （1） in, μ ∈(3.5699, 4) and x t ∈(0, 1); Step 4: Low frequency scrambling: Use formula (2) to X Sort in ascending order, [ X 2, X 1]= sort ( X )， （2） in, sort (·) is the ascending sorting function, X 1 is the index value sequence, X 2 is the sorted sequence; using X 1 and formula (3) V The elements in the array are shuffled. v 1( i )= V [ x 1( i )]， i =1, 2, …, mnk / 4， （3） in, x 1( i )∈ X 1 and V 1={ v 1( i )} is the low-frequency part after scrambling; V 1 and C The high frequency part of 1 is combined to obtain the frequency domain matrix C 2; Step 5: Inverse lifting wavelet transform: C 2 Perform the lifting wavelet inverse transform to obtain the scrambled time domain cube C 3; Step 6: Generate dynamic stereo S-box: Use formula (4) to Y Sort in ascending order, [ Y 2, Y 1]= sort ( Y )， （4） in, Y 1 is the index value sequence, Y 2 is the sorted sequence; using formula (5) Y 1 is converted into a 8×8×4 three-dimensional matrix, i.e. a dynamic stereo S-box; S = reshape ( Y 1, 8, 8, 4)， （5） in, reshape (·) is the reshaping function, S It is an 8×8×4 S-box; Step 7: Cube Diffusion: C Each pixel value in 3 d Represented in 8-bit binary d=d 7 d 6 d 5 d 4 d 3 d 2 d 1 d 0 , d l ∈{0,1}, l= 0, 1, …, 7, let x=d 7 d 6 d 5 , y=d 4 d 3 d 2 , z=d 1 d 0 , by using x , y , z The decimal value of the query element in the S box is used to replace the value of the element C Pixel value in 3 d , after replacing all pixels, we can get the encrypted image cube C 4; Step 8: Cube Decomposition: C 4 decomposes into k The size is m × n Encrypted image of J 1 , J 2 , …, J k .

Images