Multi-image encryption method based on superpixel

Abstract

A multi-image encryption method based on superpixels belongs to the field of information security. Digital image information is easily leaked or tampered during Internet transmission, and the potential security risks are becoming increasingly serious. In order to ensure the safe and reliable transmission of multiple images on a network platform, the present invention proposes a multi-image encryption method based on superpixels. The method adopts the classic framework of scrambling and diffusion: in the scrambling stage, ordinary pixel-level scrambling and bit-level scrambling are adopted; in the diffusion stage, pixel plane decomposition and binary combination are used to expand the pixel depth of the image from 8 bits to 8 k Bits are combined to form superpixels, and then superpixel diffusion is performed to obtain the encrypted image. Experiments show that this method can encrypt multiple images at the same time, with good encryption effect, high security and fast encryption speed.

Description

Technical Field

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

[0002] Nowadays, a large number of digital images are generated in many fields, such as military, medicine, national security, education and personal affairs. At the same time, a huge amount of digital image information is transmitted on the Internet every moment. Since images often contain secrets and personal privacy, they are easily attacked illegally by attackers when transmitted on the network, resulting in the leakage and tampering of such information, which faces huge threats. Therefore, image encryption has become a hot issue in academic research and practical applications.

[0003] Digital images have some unique features, such as large capacity, strong correlation between adjacent pixels, and high redundancy. In order to ensure the security of image information, a large number of image encryption algorithms have been proposed. In most encryption methods, there are often problems such as large amount of calculation, low encryption efficiency and weak security. Therefore, in order to protect the content security of multiple images, it is of great significance to study the encryption method of multiple images. The proposed multi-image encryption method based on superpixels combines the bit planes and pixel depths of multiple images to achieve the effect of superpixels. This method uses superpixels to improve the efficiency and security of the encryption method. Summary of the invention

[0004] The purpose of the present invention is to propose a multi-image encryption method based on superpixels to address the problems of low encryption efficiency and weak security of existing multi-image encryption methods.

[0005] Technical solution of the present invention: To achieve the above-mentioned invention object, the technical solution adopted is a multi-image encryption method based on superpixels, assuming that the sender is Alice and the receiver is Bob. Alice's encryption steps are detailed as follows:

[0006] Step 1: Generate pixel vector: Let k The original image is I 1 , I 2 , … , I k , whose sizes are m × n , whose pixel value is (0, 255) and pixel depth is 8 bits, I 1 , I 2 , … , I k Convert to k The length is mn , and transform this kThe vectors are connected end to end to form a vector with a length of kmn One-dimensional vector of S 1 ;

[0007] Step 2: Generate a chaotic sequence: Randomly select the initial value of the one-dimensional Sin-Tent map shown in formula (1) x 0 ∈(0, 1) and control parameters μ 1 , μ 1 For any real number, iterate the mapping kmn times, we can get a length of kmn Chaotic sequence X 1 ; Randomly select the initial value of the one-dimensional Sin-Tent mapping shown in formula (1) y 0 ∈(0, 1) and control parameters μ 2 , μ 2 For any real number, iterate the mapping 8 kmn times, we get a length of 8 kmn Chaotic sequence Y 1 ,

[0008] ; (1)

[0009] Step 3: Ordinary pixel-level scrambling: X 1 Arrange in ascending order to get the index value vector W ,use W right S 1 Sorting and scrambling can get a length of kmn One-dimensional vector of S 2 ,

[0010] S 2 ( r )= S 1 ( W ( r ), r =1, 2, …, kmn ; (2)

[0011] Step 4: Binary conversion: Convert the matrix S 2 Convert each pixel in the image into binary and connect them in sequence to get a pixel with a length of 8. kmn One-dimensional vector of S 3 ;

[0012] Step 5: Bit-level scrambling: Y 1 Arrange in descending order to get the index value vector H ,use H right S 3 Sorting and scrambling can get a length of 8 kmn One-dimensional vector of S 4 ;

[0013] Step 6: Generate superpixels: S 4 Each of the 8 consecutive k The bits are grouped together and converted into a decimal number to obtain a number with a length of mn One-dimensional vector of S 5 , S 5 The pixels in are called superpixels;

[0014] Step 7: Chaos integerization: Select Y 1 Before mn Elements constitute a chaotic sequence Y 2 ,right Y 2 After integer processing, we can get a length of mn Integer chaotic sequence Y 3 ,

[0015] Y 3 =mod( floor ( Y 2 ×10 8 ), 2 8k ), (3)

[0016] in, floor (·) is the floor function, mod(·) is the modulus operation function;

[0017] Step 8: Superpixel Diffusion: Utilization Y 3 ,right S 5 Perform the following superpixel diffusion operation to obtain a superpixel with a length of mn One-dimensional vector S 6 ,

[0018] S 6 (1)= S 5 (1) Y 3 (1), (4)

[0019] S 6 ( j )= S 5 ( j )⊕ Y 3 ( j )⊕ S 6 ( j -1), j >1, (5)

[0020] Among them, ⊕ is the exclusive OR operation;

[0021] Step 9: Generate a hyperimage: S 6 Convert to a size of m × n Super Image C , which is the encrypted image, with a pixel depth of 8 k Bit.

[0022] Furthermore, in step 6, the value of the superpixel is in the interval [0, 2 8k -1], that is, the pixel depth of the superpixel is 8 k Bit.

[0023] Furthermore, in step 9, the super-image is an image composed of super-pixels.

[0024] Bob’s decryption process is the reverse process of Alice’s encryption process. In the decryption process, the same chaotic sequence is used to encrypt the encrypted image. C Decryption operation can be performed to recover k The original image.

[0025] Beneficial effects: Aiming at the shortcomings of existing encryption methods, such as large amount of calculation, low encryption efficiency and weak security, the present invention proposes a multi-image encryption method based on superpixels. The main contributions are as follows: (1) A multi-image encryption method based on superpixels is proposed; (2) The concepts of superpixels and superimages are defined, and the binary matrices of multiple images are superimposed together and converted into binary to form superpixels, thereby merging multiple images into a superimage; (3) The proposed multi-image encryption method has the characteristics of high efficiency and strong security, and can realize the secure transmission of multiple images. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Picture 1 : Flowchart of multi-image encryption based on superpixels;

[0027] Picture 2 : original image set;

[0028] Picture 3 : Schematic diagram of superpixel formation;

[0029] Picture 4 : Encrypted image. DETAILED DESCRIPTION

[0030] The embodiments of the present invention are further described in detail below with reference to specific drawings and examples.

[0031] Picture 1 This is a multi-image encryption flow chart based on superpixels. The programming software used is Matlab R2016b. Picture 2 The four gray images with a size of 512×512 are shown as original images. Using the multi-image encryption method based on superpixels, Alice's encryption process is detailed as follows.

[0032] Step 1: Generate chaotic sequence: read 4 gray images of size 512×512, whose pixel values ​​are (0,255) and pixel depth is 8 bits, and convert these 4 images into 4 one-dimensional vectors of length 262144 respectively, and connect these 4 one-dimensional vectors at the first position to form a one-dimensional vector of length 1048576 S 1 ;

[0033] Step 2: Generate a chaotic sequence: Select the initial value of the one-dimensional Sin-Tent map shown in formula (1) x 0 =0.385421 and control parameters μ 1 =3.1415926, iterating the mapping 1048576 times, we can get a chaotic sequence with a length of 1048576 X 1 ; Select the initial value of the one-dimensional Sin-Tent mapping shown in formula (1) y 0 =0.156565 and control parameters μ 2 =3.756245, iterating the mapping 8388608 times, we can get a chaotic sequence with a length of 8388608 Y 1 ;

[0034] Step 3: Ordinary pixel-level scrambling:X 1 Arrange in ascending order to obtain its index value vector W ,right S 1 use W Sorting and scrambling with formula (2) can obtain a one-dimensional vector with a length of 1048576 S 2 ;

[0035] Step 4: Binary conversion: Convert the matrix S 2 Each pixel in is converted into binary and connected in sequence to obtain a one-dimensional vector with a length of 8388608 S 3 ;

[0036] Step 5: Bit-level scrambling: Y 1 Sort in descending order to get its index value vector H ,use H right S 3 Sorting and scrambling can get a one-dimensional vector with a length of 8388608 S 4 ;

[0037] Step 6: Generate superpixels: S 4 Each of the 8 consecutive k The bits are grouped together and converted into decimal numbers, resulting in a one-dimensional vector of length 262144 S 5 , S 5 The pixels in are called superpixels. Taking 4 pixels 168, 30, 0 and 30 as an example, the number of superpixels is 2820538398. The specific construction process is as follows: Picture 2 As shown;

[0038] Step 7: Chaos integerization: Select Y 1 The first 262144 elements of form a chaotic sequence Y 2 ,right Y 2 After integer processing, we can get an integer chaotic sequence with a length of 262144. Y 3 ;

[0039] Step 8: Superpixel Diffusion: Utilization Y 3 And formula (4) and (5) S 5Performing superpixel diffusion operation, we can get a one-dimensional vector with a length of 262144 S 6 ;

[0040] Step 9: Generate a hyperimage: S 6 Converted into a superimage of size 512×512 C , which is an encrypted image with a pixel depth of 32 bits. Picture 4 shown.

[0041] Bob’s decryption process is the reverse process of Alice’s encryption process. In the decryption process, the same chaotic sequence is used to decrypt the encrypted image, and the four original images can be restored. Picture 3 shown.

Claims

1. A multi-image encryption method based on superpixels, characterized in that: The encryption process includes the following steps: Step 1: Generate pixel vector: Let k The original image is I 1, I 2, … , I k , whose sizes are m × n , whose pixel value is (0, 255) and pixel depth is 8 bits, I 1, I 2, … , I k Convert to k The length is mn , and transform this k The vectors are connected end to end to form a vector with a length of kmn One-dimensional vector of S 1; Step 2: Generate a chaotic sequence: Randomly select the initial value of the one-dimensional Sin-Tent map shown in formula (1) x 0∈(0, 1) and control parameters μ 1, μ 1 is any real number, iterate the mapping kmn times, we can get a length of kmn Chaotic sequence X 1; Randomly select the initial value of the one-dimensional Sin-Tent mapping shown in formula (1) y 0∈(0, 1) and control parameters μ 2, μ 2 is any real number, iterate the mapping 8 kmn times, we get a length of 8 kmn Chaotic sequence Y 1, ； （1） Step 3: Ordinary pixel-level scrambling: X 1 is sorted in ascending order to obtain its index value vector W ,use W right S 1 is sorted and shuffled to obtain a length of kmn One-dimensional vector of S 2, S 2( r )= S 1( W ( r ))， r =1, 2, …, kmn ； （2） Step 4: Binary conversion: Convert the matrix S Each pixel in 2 is converted into binary and connected in sequence to obtain a length of 8 kmn One-dimensional vector of S 3; Step 5: Bit-level scrambling: Y 1 is sorted in descending order to obtain its index value vector H ,use H right S 3 sort and shuffle to get a length of 8 kmn One-dimensional vector of S 4; Step 6: Generate superpixels: S 4 of each consecutive 8 k The bits are grouped together and converted into a decimal number to obtain a number with a length of mn One-dimensional vector of S 5. S The pixels in 5 are called superpixels; Step 7: Chaos integerization: Select Y 1 ago mn Elements constitute a chaotic sequence Y 2. Yes Y 2 is integerized to get a length of mn Integer chaotic sequence Y 3. Y 3=mode( floor ( Y 2×10 8 ), 2 8k ), (3) in, floor (·) is the floor function, mod(·) is the modulus operation function; Step 8: Superpixel Diffusion: Utilization Y 3. Yes S 5 Perform the following superpixel diffusion operation to obtain a superpixel with a length of mn One-dimensional vector S 6. S 6(1)= S 5(1)⊕ Y 3(1)， （4） S 6( j )= S 5( j )⊕ Y 3( j )⊕ S 6( j -1)， j >1， （5） Among them, ⊕ is the exclusive OR operation; Step 9: Generate a hyperimage: S 6 is converted into a size of m × n Super Image C , which is the encrypted image, with a pixel depth of 8 k Bit.

2. The method according to claim 1, characterized in that: In step 6, the value of the superpixel is in the interval [0,2 8k -1], that is, the pixel depth of the superpixel is 8 k Bit.

3. The method according to claim 1, characterized in that: In step 9, the super-image is an image composed of super-pixels.

Images