CN112182543B

CN112182543B - Visual password method

Info

Publication number: CN112182543B
Application number: CN202011098265.1A
Authority: CN
Inventors: 王晶; 王凌福; 黄位甲
Original assignee: Guilin University of Electronic Technology
Current assignee: Guilin University of Electronic Technology
Priority date: 2020-10-14
Filing date: 2020-10-14
Publication date: 2024-04-30
Anticipated expiration: 2040-10-14
Also published as: CN112182543A

Abstract

The invention discloses a visual cryptography method, which comprises the steps of firstly acquiring basic parameters, simultaneously converting an acquired original picture into a gray level picture matrix, then initializing a third-order tensor to store a base matrix set, simultaneously generating a first matrix subject to uniform distribution, traversing the first matrix, assigning values to the third-order tensor by utilizing the values of the first matrix to obtain a corresponding base matrix set, then generating a corresponding noise matrix set according to the acquired security parameters, finally generating a component matrix according to the gray level picture matrix, the base matrix and the noise matrix, and converting the component matrix into a component picture to finish encryption of the picture; for component diagram restoration, a plurality of component diagrams are converted into corresponding component matrixes, and modulo addition operation is carried out, so that a restored picture is obtained, a non-binarized image can be directly processed, and the method can be further popularized to the processing of a color image, and the flexibility and the expandability are improved.

Description

Visual password method

Technical Field

The invention relates to the technical field of visual cryptography, in particular to a visual cryptography method.

Background

The "visual cryptography" is based on a threshold secret sharing concept, combining secret sharing with digital images. The visual cryptography shares the original image in several transparent films called component images, each of which does not reveal any information of the original image. The original image can be directly restored by only superposing films together during decryption. However, most of the existing visual cryptography is directed at a binary image, that is, the gray value of a pixel point in the image is 0 or 255, the binary image carries little information, and the application range of the binary image is small in real life, so that most of the existing visual cryptography cannot be implemented based on a general image, is applicable to a scene, has a limited range, and cannot be widely applied. In addition, most schemes need to expand the original image, namely the dimensions of the component image and the restored image are different from those of the original image, which increases storage redundancy and reduces operation efficiency. In addition, many schemes use a fixed base matrix, and flexibility and expandability are required to be improved.

Disclosure of Invention

The invention aims to provide a visual cryptography method which improves flexibility and expandability.

In order to achieve the above object, the present invention provides a visual cryptography method comprising the steps of:

Basic parameters are acquired, and the acquired original picture is converted into a gray scale image matrix;

initializing a third-order tensor, and obtaining a base matrix set according to a first matrix generated by traversal;

generating a noise matrix set according to the acquired safety parameters;

generating a component matrix according to the gray map matrix, the base matrix and the noise matrix, and converting the component matrix into a component map;

And converting the plurality of component images into corresponding component matrixes, performing modulo addition operation, and simultaneously storing and converting to obtain a restored image.

Initializing a third-order tensor, and obtaining a base matrix set according to a first matrix generated by traversing, wherein the method comprises the following steps:

initializing a third-order tensor to store a base matrix set, generating a first matrix obeying uniform distribution, traversing the first matrix, and assigning values to the third-order tensor by using the values of the first matrix to obtain the corresponding base matrix set.

Generating a component matrix according to the gray map matrix, the base matrix and the noise matrix, and converting the component matrix into a component map, wherein the method comprises the following steps:

And calculating the Hadamard product of the base matrix and the gray map matrix, summing the Hadamard product with the noise matrix, and then carrying out a residual operation to obtain a component matrix.

Generating a component matrix according to the gray map matrix, the base matrix and the noise matrix, converting the component matrix into a component map, and further comprising:

and storing any one value of RGB values in the gray scale map matrix for three times, and converting the obtained third-order tensor into a component map.

Converting the plurality of component graphs into corresponding component matrixes, performing modulo addition operation, and simultaneously storing and converting to obtain a restored picture, wherein the method comprises the following steps of:

and obtaining a plurality of component graphs, converting the component graphs into a plurality of corresponding component matrixes, performing modular addition operation on the plurality of obtained component matrixes, converting the obtained gray-scale graph matrixes into restored pictures, and storing the restored pictures.

The invention relates to a visual cryptography method, which comprises the steps of firstly obtaining basic parameters, simultaneously converting an obtained original picture into a gray level picture matrix, then initializing a third-order tensor to store a basic matrix set, simultaneously generating a first matrix subject to uniform distribution, traversing the first matrix, assigning values to the third-order tensor by utilizing the values of the first matrix to obtain a corresponding basic matrix set, then generating a corresponding noise matrix set according to the obtained safety parameters, finally generating a component matrix according to the gray level picture matrix, the basic matrix and the noise matrix, and converting the component matrix into a component picture to finish encryption of the picture; for component diagram restoration, a plurality of component diagrams are converted into corresponding component matrixes, and modulo addition operation is carried out, so that a restored picture is obtained, a non-binarized image can be directly processed, and the method can be further popularized to the processing of a color image, and the flexibility and the expandability are improved.

Drawings

In order to more clearly illustrate the embodiments of the invention or the technical solutions in the prior art, the drawings that are required in the embodiments or the description of the prior art will be briefly described, it being obvious that the drawings in the following description are only some embodiments of the invention, and that other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.

Fig. 1 is a schematic diagram of steps of a visual cryptography method according to the present invention.

Fig. 2 is a schematic flow chart of image encryption provided by the invention.

Fig. 3 is a schematic flow chart of image restoration provided by the invention.

Fig. 4 is a flowchart of the generation of a base matrix set provided by the present invention.

FIG. 5 is a graph of peak signal to noise ratio of a picture restored by using different numbers of encryption component graphs and an original picture.

Fig. 6 is a structural similarity line diagram of a picture restored by using different numbers of encryption component diagrams and an original picture.

Fig. 7 is a peak signal to noise ratio line diagram of an encryption component diagram and an original picture, which are provided by the invention and added with Laplace noise with different sizes.

Fig. 8 is a structural similarity line diagram of an encryption component diagram added with Laplace noise with different sizes and an original picture.

Detailed Description

Embodiments of the present invention are described in detail below, examples of which are illustrated in the accompanying drawings, wherein like or similar reference numerals refer to like or similar elements or elements having like or similar functions throughout. The embodiments described below by referring to the drawings are illustrative and intended to explain the present invention and should not be construed as limiting the invention.

In the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.

Referring to fig. 1, the present invention provides a visual cryptography method, which includes the following steps:

s101, acquiring basic parameters, and converting the acquired original picture into a gray scale image matrix.

Specifically, firstly, parameters are determined, the number N of component images, a restoration threshold t, a safety parameter epsilon and the dimension m multiplied by N of an original image are selected, for example, the number N of generated component images is selected to be 50, the restoration threshold t is 30, the privacy budget epsilon is 3, and the dimension of the original image is 512 multiplied by 512; for accurate manipulation of the picture, the picture can be regarded as a third-order tensor, which is a long and wide channel and an RGB channel, respectively, and since the RGB three values in the gray map are the same, the gray map can be abstracted into a matrix. For example, the original picture used in this example is a gray-scale picture of 512×512, the RGB three-channel values of each pixel are the same, and the original picture can be converted into the gray-scale picture matrix M by only storing one RGB value.

S102, initializing a third-order tensor, and obtaining a base matrix set according to the first matrix generated by traversing.

Specifically, the following conditions are required to be satisfied for generating the base matrix: firstly, the base matrices are the same as the original picture gray map matrix in dimension, secondly, the sum of the 0-1 base matrices should be an all-one matrix, and finally, the 1 positions in the 0-1 base matrices are not overlapped.

1. A third order tensor e= (E _kij)_N×m×n,K represents the dimension of the number of component graphs, i represents the length of the component graphs, and j represents the width of the component graphs. Wherein the third-order tensor is used to preserve the set of basis matrices.

2. A first matrix a= (a _ij)_m×n, and a _ij -U [0,1] is generated, indicating that a _ij obeys a uniform distribution over [0,1 ].

3. Traversing the first matrix A, letting e _{k ij＝1} if and only ifK is more than or equal to 0 and less than or equal to N-2 or

And assigning values to the initialized third-order tensors according to the values of the first matrix to obtain corresponding base matrix sets, wherein the initialized third-order tensors are all 0, traversing the first matrix according to the judging condition, and setting part 0 in the third-order tensors as 1 to obtain 0-1 base matrix sets.

The dimensions of the base matrix are 512×512 in this example, the sum of the 0-1 base matrices should be an all-one matrix, and finally the positions of 1 in the 0-1 base matrices are non-overlapping, and the flow chart of the base matrix generation algorithm is shown in fig. 4. Taking 5×5 all-one matrix decomposition into 3 base matrices as an example, the conditions that the base matrix needs to satisfy are illustrated:

next, a practical example of how to generate a base matrix that meets the conditions is described, which is to generate 50 base matrices with dimensions 512×512.

1. A third order tensor e= (E _kij)_50×512×512,

2. A matrix a= (a _ij)_512×512, and a _ij to U [0,1]. A base matrix set is generated from the values of the matrix.

3. Traversing matrix A, let e _{k ij＝1}, if and only ifK is more than or equal to 0 and less than or equal to 48 or/>

S103, generating a noise matrix set according to the acquired safety parameters.

Specifically, according to the selected safety parameter epsilon, a noise matrix set R= (R _kij)_N×m×m(r_kij -Lap (255/(3×30)) which obeys Laplace distribution is generated, and N m×n-order noise matrices which obey Laplace distribution are obtained. According to the above example, the noise matrix set is r= (R _kij)_50×512×512(r_kij -Lap (255/(3×30))), and 50 512×512-order noise matrices obeying the Laplace distribution are obtained.

S104, generating a component matrix according to the gray map matrix, the base matrix and the noise matrix, and converting the component matrix into a component map.

Specifically, the steps of generating the component matrix by the gray map matrix, the base matrix and the noise matrix are as follows:

S_i＝(E_i⊙M+R_k)mod 256 i≤N

Where S _i represents the resulting component matrix, E _i represents the base matrix, M represents the gray map matrix of the original picture, R _k is the noise matrix following the Laplace distribution, R _k＝(r_kij)_m×n. The ". Iy represents Hadamard product. An m×n order matrix a= (a _ij)_m×n,B＝(b_ij)_m×n, let c=a=b= (C _ij)_m×n, C _ij＝a_ij×b_ij is present, i is less than or equal to m, j is less than or equal to N. Get a component matrix set s= { S _i |i is less than or equal to N }.

Taking a 3X 3 matrix as an example, the definition of the matrix Hadamard product is better understood.

The conversion of the component matrix into the component map is similar to the conversion of the third-order tensor into the matrix, and RGB three values in the gray map are the same, so that the matrix can be converted into the third-order tensor by only storing one value in RGB three times, and finally the third-order tensor is converted into the component map. The generated 50 component images are stored and distributed to 50 different participants to complete the encryption of the images, so that the non-binarized images can be directly processed and further popularized to the processing of color images.

S105, converting the plurality of component graphs into corresponding component matrixes, performing modulo addition operation, and simultaneously storing and converting to obtain a restored picture.

Specifically, first, a map T (T. Gtoreq.t) Zhang Fenliang is collected, in this example, T is 30, that is, at least 30 maps Zhang Fenliang are collected, and these 30 component maps are converted into 30 corresponding gray map matrices according to the previous method.

Secondly, the restoring process only needs to carry out modulo addition operation on gray scale image matrixes corresponding to the 30 component images, namely:

M'＝(∑S_i')mod 256 0≤i≤29

Wherein M 'is a gray scale map matrix corresponding to the reduction map, S _i' belongs to S, and is a collected component matrix.

Finally, the same method as the above method is used for converting the gray image matrix into a restored image and storing the restored image to complete the image restoration process.

Referring to fig. 2 and 3, the encryption and recovery processes are mainly included. The process of picture encryption includes selecting appropriate parameters: and generating the number N of the component graphs, restoring a threshold t, a safety parameter epsilon and the dimension m multiplied by N of the original picture. N0-1 base matrixes are generated, the dimensions of the base matrixes are the same as those of the gray-scale image matrixes corresponding to the original pictures, the 0-1 base matrixes are required to be all-one matrixes, and the positions of 1 in the 0-1 base matrixes are not overlapped. And then multiplying the 0-1 base matrixes by gray image matrixes corresponding to the original pictures respectively, and adding the multiplied matrixes and noise matrixes obeying Laplace distribution to generate a plurality of component matrixes. And finally, converting the component matrix into a component diagram. The basic requirement of the component map is that the related information of the original picture cannot be acquired by naked eyes. And then, the original picture can be restored by utilizing a certain number of component pictures, so that the restored picture can be identified by naked eyes, and the related information of the original picture can be basically obtained. The reduction process is as follows: firstly, a certain number of component graphs are collected and converted into component matrixes, then the matrixes are subjected to modulo addition operation, and finally, the summation matrixes are converted into a reduction graph. The invention ensures that the information of the original picture is protected by encrypting the original picture, and simultaneously, the information is restored by the component diagram, and the related information of the original picture can be obtained.

According to analysis of the restored images obtained by restoring 10, 20, 30, 40 and 50 component images, it can be seen that most of pixels of the original image are concentrated between 25 and 225, the distribution is uniform, and most of pixels of the component image are concentrated at two ends, between 0 and 25 and 225 and 255, and the distribution is non-uniform. As the number of component maps for restoration increases, the distribution of pixel values gets closer to the middle, closer to the pixel distribution of the original map.

In order to prove the effect of the patent, experimental effect graphs are shown in fig. 5, 6, 7 and 8, and are respectively the comparison between the original picture and the reducing picture and the comparison between the original picture and the component picture, so that the effect of encryption and reduction can be measured. FIG. 5 is a graph of peak SNR for a picture restored by using different numbers of component graphs versus an original picture. FIG. 6 is a structural similarity plot of a restored picture with a different number of component maps to an original picture. The abscissa is the number of component graphs, and the ordinate is the peak signal-to-noise ratio (PNSR) or Structural Similarity (SSIM), the greater the peak signal-to-noise ratio (PNSR) or Structural Similarity (SSIM), the higher the degree of similarity of the reduced graph to the original graph. The different curves represent the lambda size of the added Laplace random value, the larger lambda represents the larger added noise, the larger the interference on the image is, the broken line represents the PNSR and SSIM values of the basically identifiable reducing image for the reference line, and when the value is lower than the reference line, the human eyes cannot identify the reducing image, namely the related information of the original image cannot be acquired.

Fig. 7 is a peak signal to noise ratio plot of a component plot of added Laplace noise of different magnitudes versus an original picture. Fig. 8 is a structural similarity plot of the component diagram with added Laplace noise of different magnitudes and the original picture. The abscissa represents lambda of the added Laplace random value, the ordinate represents peak signal to noise ratio (PNSR) or Structural Similarity (SSIM) of the component images and the original image, and different curves represent the quantity N of the component images generated by encrypting one original image, wherein the larger the N is, the less original image information is carried by each component image. The broken line represents the PNSR, SSIM values that cannot identify the component map, and when the value is lower than the reference line, the human eye cannot identify the component map, that is, the encryption effect can be achieved.

The invention relates to a visual cryptography method, which comprises the steps of firstly acquiring basic parameters, simultaneously converting an acquired original picture into a gray level picture matrix, then initializing a third-order tensor to store a base matrix set, simultaneously generating a first matrix subject to uniform distribution, traversing the first matrix, assigning values to the third-order tensor by utilizing the values of the first matrix to obtain a corresponding base matrix set, then generating a corresponding noise matrix set according to the acquired safety parameters, finally generating a component matrix according to the gray level picture matrix, the base matrix and the noise matrix, and converting the component matrix into a component picture to finish encryption of the picture; for component diagram restoration, a plurality of component diagrams are converted into corresponding component matrixes, and modulo addition operation is carried out, so that a restored picture is obtained, a non-binarized image can be directly processed, and the method can be further popularized to the processing of a color image, and the flexibility and the expandability are improved.

The above disclosure is only a preferred embodiment of the present invention, and it should be understood that the scope of the invention is not limited thereto, and those skilled in the art will appreciate that all or part of the procedures described above can be performed according to the equivalent changes of the claims, and still fall within the scope of the present invention.

Claims

1. A visual cryptography method comprising the steps of:

generating a noise matrix set according to the acquired safety parameters;

Converting the plurality of component images into corresponding component matrixes, performing modulo addition operation, and simultaneously storing and converting to obtain a restored image;

Initializing a third-order tensor, and obtaining a base matrix set according to a matrix generated by traversing, wherein the method comprises the following steps:

Initializing a third-order tensor to store a base matrix set, generating a first matrix obeying uniform distribution, traversing the first matrix, and assigning values to the third-order tensor by utilizing the values of the first matrix to obtain the corresponding base matrix set;

generating a noise matrix set according to the acquired safety parameters, including:

generating a noise matrix set obeying Laplace distribution according to the selected safety parameters;

generating a component matrix according to the gray map matrix, the base matrix and the noise matrix, and converting the component matrix into a component map, wherein the method comprises the following steps of:

Calculating the Hadamard product of a base matrix and the gray map matrix, summing the Hadamard product with the noise matrix, and then carrying out a residual operation to obtain a component matrix;

Storing any one value of RGB values in the gray scale map matrix for three times, and converting the obtained third-order tensor into a component map;