WO2020029334A1 - 一种灰度图像可视加密方法 - Google Patents
一种灰度图像可视加密方法 Download PDFInfo
- Publication number
- WO2020029334A1 WO2020029334A1 PCT/CN2018/102211 CN2018102211W WO2020029334A1 WO 2020029334 A1 WO2020029334 A1 WO 2020029334A1 CN 2018102211 W CN2018102211 W CN 2018102211W WO 2020029334 A1 WO2020029334 A1 WO 2020029334A1
- Authority
- WO
- WIPO (PCT)
- Prior art keywords
- image
- image block
- pixel
- visual
- error
- Prior art date
Links
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/08—Key distribution or management, e.g. generation, sharing or updating, of cryptographic keys or passwords
- H04L9/0816—Key establishment, i.e. cryptographic processes or cryptographic protocols whereby a shared secret becomes available to two or more parties, for subsequent use
- H04L9/0838—Key agreement, i.e. key establishment technique in which a shared key is derived by parties as a function of information contributed by, or associated with, each of these
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/08—Key distribution or management, e.g. generation, sharing or updating, of cryptographic keys or passwords
- H04L9/0816—Key establishment, i.e. cryptographic processes or cryptographic protocols whereby a shared secret becomes available to two or more parties, for subsequent use
- H04L9/085—Secret sharing or secret splitting, e.g. threshold schemes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T1/00—General purpose image data processing
- G06T1/0021—Image watermarking
- G06T1/0028—Adaptive watermarking, e.g. Human Visual System [HVS]-based watermarking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/10—Segmentation; Edge detection
- G06T7/11—Region-based segmentation
-
- G—PHYSICS
- G09—EDUCATION; CRYPTOGRAPHY; DISPLAY; ADVERTISING; SEALS
- G09C—CIPHERING OR DECIPHERING APPARATUS FOR CRYPTOGRAPHIC OR OTHER PURPOSES INVOLVING THE NEED FOR SECRECY
- G09C5/00—Ciphering apparatus or methods not provided for in the preceding groups, e.g. involving the concealment or deformation of graphic data such as designs, written or printed messages
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L9/00—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols
- H04L9/06—Cryptographic mechanisms or cryptographic arrangements for secret or secure communications; Network security protocols the encryption apparatus using shift registers or memories for block-wise or stream coding, e.g. DES systems or RC4; Hash functions; Pseudorandom sequence generators
Definitions
- the invention relates to the field of multimedia communication security, in particular to a method for visually encrypting grayscale images.
- Visual encryption is an efficient secret sharing method that does not require any calculation during the decoding process.
- a (n, k) threshold visual password it divides a secret image into n parts and only obtains at least k of them. Overlay it to recover the content of the secret image.
- research on visual cryptography has two focuses: one is to reduce the size of the stored image, and the other is to improve the visual quality of the reconstructed image. Among them, the size of the shared image is generally expanded relative to the size of the secret image.
- a small expansion factor or even a non-expandable visual password has many advantages, such as low processing complexity, small storage space, and low transmission bandwidth.
- typical size-invariant visual cryptographic algorithms include random grid method, probability algorithm, and block coding method.
- the quality of the reconstructed image is greatly reduced.
- gray-scale secret images the quality of the reconstructed images is far from being satisfactory.
- the secret images of gray and color image visual passwords are gray and color images, and the reconstructed image is a half-tone-like image.
- the current research on Size-Invariant Visual Cryptography can be roughly divided into two categories. One is the scheme of separating halftone and visual cryptography, and the other is the scheme of direct block quantization and mapping. among them:
- the first category is about the separation of halftone and visual cryptography:
- the grayscale image is first converted into a binary image, for example, using halftoning or a simple fixed-threshold binary quantization technique, and then visually encoding the binary image.
- This solution separates the halftone process from the visual encryption process: the halftone process does not consider subsequent visual password encryption; the visual encryption does not consider whether the processed image is halftone, or two similar to the company logo Value image.
- This type of algorithm is currently the mainstream of visual cryptography for grayscale images.
- the second type is about direct block quantization and mapping:
- the lossy mapping from the secret pixel / image block to the reconstructed pixel / image block is not compensated by other surrounding pixels and image blocks. Due to the limitation of the security conditions of visual cryptography, the ability to express secret binary images on reconstructed images is limited. For example, if a 2 by 2 image block is used, there are 5 types of image blocks for the secret image according to the number of black pixels in the block. However, in the reconstructed image, only 3 possible image blocks are allowed, of which the number of black pixels They are 2, 3 and 4. In this way, there is a many-to-one lossy mapping from the secret image block to the reconstructed image block. None of the current algorithms can compensate for this loss of brightness.
- the existing algorithms do not directly optimize the perceived quality of the reconstructed image.
- the image quality measures used by the existing algorithms do not reflect how the human visual system perceives halftone images.
- the spectral characteristics of the reconstructed image are not reflected in existing quality measures.
- the quality measures commonly used in the existing literature are intra-block variance and global contrast.
- the intra-block variance measure measures the uniformity of the reconstructed image relative to the secret binary image. However, it does not measure the perceived quality of the halftone reconstructed image.
- the global contrast measure measures the brightness difference between image blocks used to represent black secret pixels and white secret pixels on the reconstructed image. This measure is designed for binary secret images. Recently, XTWu et al. Proposed a visual password for blue noise-shared images, but after superimposing two blue noise-shared images, the reconstructed image will no longer have blue noise characteristics, so this type of algorithm is only suitable for gray Degree secret images are simply binarized and then encrypted, so they have very low detail resolution.
- the purpose of the present invention is to provide a method for visually encrypting a grayscale image to improve the visual quality of a reconstructed target image.
- the present invention adopts the following technical solutions:
- a grayscale image visual encryption method includes the following steps:
- the image after color gamut mapping is divided into blocks and divided into multiple non-overlapping image blocks of size B ⁇ B;
- a vector quantizer is used to quantize it into a binary image block, that is, each pixel value is 0 or 1 after quantization;
- the quantized binary image block is subjected to vector visual cryptographic coding to obtain k divided image blocks s 1 [n], ..., sk [n];
- the error after diffusion is Use error Modify the input pixel values of the above four image blocks to get the modified image blocks This modified image patch As the input of the vector quantizer in step s13;
- the step s11 is specifically:
- x [n] is the image after color gamut mapping.
- the step s13 is specifically:
- Each pixel is quantized using a binary scalar quantizer, the quantization threshold is set to 1/2, and the output reconstruction points are set to 0 and 1.
- the step s14 is specifically:
- D is the set of all possible white pixels in the target image block,
- the columns of the matrix C j are scrambled, and each row of the matrix C j is rearranged into B ⁇ B image blocks, and assigned to a sub-image as the corresponding image block of the sub-image, thus obtaining k sub-image blocks s 1 [n], ..., s k [n].
- step s142 the selection method of the matrix C j is:
- the step s17 is specifically:
- the pixel value will not be changed.
- the pixel will keep the diffused error intact and use the same scalar error diffusion filter to pass it to the pixels around it.
- the method is applicable to any of the following three types of visual passwords designed for binary images, namely:
- Probabilistic visual password random lattice visual password, and vector visual password.
- the method of the present invention can be easily combined with any of the existing probability visual passwords for binary images, random lattice visual passwords, and vector visual passwords to construct visual passwords for grayscale images. , And can significantly improve the visual quality of the reconstructed target image.
- the method of the present invention satisfies the contrast and security conditions of a gray-scale visual password. As long as the basic binary visual password is perfect and secure, the gray-scale image visual password obtained by this method is perfect and secure.
- FIG. 1 is a principle block diagram of a grayscale image visible encryption method according to an embodiment of the present invention.
- FIG. 2 is a flowchart of a method for visually encrypting a grayscale image according to an embodiment of the present invention.
- FIG. 3 is another flowchart block diagram of a grayscale image visible encryption method according to an embodiment of the present invention.
- FIG. 4 is a schematic diagram of vector quantization using scalar quantization and intra-block error diffusion in an embodiment of the present invention.
- FIG. 5 is a schematic diagram of vector error diffusion in an embodiment of the present invention.
- FIG. 6 is a comparison result diagram of the method in the embodiment of the present invention and the algorithm of Wang, Yang, and others.
- an encryption-analysis-by-synthesis (AbS) framework with closed-loop feedback is proposed.
- a visual password is designed from the perspective of analysis-synthesis.
- the process of segmenting a secret image into separate images The analysis part is the synthesis part, and the reconstructed image is generated by superimposing the stored images on the encryptor.
- AbS The basic meaning of AbS is to decode the stored image obtained by encryption in the encryptor (not in the decoder), and use the decoded result as the basis for optimizing the visual quality.
- the reconstructed image synthesized by the encryptor is fed back to the analysis part to obtain the difference between the current reconstructed pixel (or image block) and the secret pixel (or image block).
- the halftone process can Carefully adjust the ratio of black and white pixels in the binary image it produces to compensate for the above differences. From a structural point of view, the visual encryption and decoding are moved into the feedback loop of the error diffusion halftone algorithm.
- the AbS framework proposed by the present invention has high flexibility, and it can be combined with three existing size-invariant visual passwords, that is, it can be combined with probabilistic visual passwords, random lattice visual passwords, and vector visual passwords.
- a probabilistic visual password based on AbS a random lattice visual password based on AbS
- a vector visual password based on AbS are obtained, respectively.
- the obtained encryption system has the perfect security defined by Shannon (Perfect Secrecy).
- a combination of the method of the present invention and a vector visual password is used as an example to explain the visual encryption process of a grayscale image:
- a grayscale image visual encryption method includes the following steps:
- the color gamut of the grayscale image J [n] is mapped to the color gamut of the visual password superimposed reconstructed image (ie, the target image).
- the step s11 is specifically:
- Maximum brightness The calculation method is as follows: exhaustively all possible target image blocks, for each image block, calculate the number of white pixels in it, which is the brightness of the image block, and calculate the maximum value of all these brightnesses, which is
- x [n] is the image after color gamut mapping.
- the color gamut mapping using the above steps can ensure the stability of the entire error diffusion feedback loop.
- the input image may be first subjected to histogram equalization, and then the above-mentioned color gamut mapping may be used.
- the image after color gamut mapping is divided into blocks and divided into a plurality of image blocks of size B ⁇ B which do not overlap each other.
- the image blocks are sequentially accessed in the order of raster scanning, and one image block is processed at a time.
- a vector quantizer For each image block, a vector quantizer is used to quantize it into a binary image block, that is, each pixel takes a value of 0 or 1 after quantization.
- the step s13 is specifically:
- Each pixel is quantized using a binary quantizer, the quantization threshold is set to 1/2, and the output reconstruction points are set to 0 and 1.
- the quantized binary image block is subjected to vector visual cryptographic coding to obtain k divided image blocks s 1 [n], ..., sk [n].
- the step s14 is specifically:
- D is the set of all possible white pixels in the target image block
- represents the number of elements in the set D.
- the selection method of the matrix C j is:
- mapping in Table 1 can be used, that is, the above conditions are satisfied.
- the columns of the matrix C j are scrambled, and each row of the matrix C j is rearranged into B ⁇ B image blocks, and assigned to a sub-image as the corresponding image block of the sub-image, thus obtaining k sub-image blocks s 1 [n], ..., s k [n].
- the error after diffusion is Use error Modify the input pixel values of the above four image blocks to get the modified image blocks This modified image patch As input to the vector quantizer in step s13.
- This step s17 is specifically:
- the four image blocks to the right, bottom left, bottom, and bottom right of the current image block are selected as the neighborhoods of the current image block, and these neighborhoods are the destinations of the current image block error diffusion.
- the ordinary scalar error diffusion filter is, for example, a commonly used Floyd-Steinberg filter.
- the pixel value will not be changed.
- the pixel will keep the diffused error intact and use the same scalar error diffusion filter to pass it to the pixels around it.
- the method of the present invention is organically combined with the vector visual password to construct a visual password for a grayscale image, and can significantly improve the visual quality of the reconstructed target image.
- a scalar quantizer is used in combination with intra-block error diffusion to achieve.
- a schematic of this implementation is shown in Figure 4.
- the slashed box indicates the quantized pixels
- the white box indicates the quantized pixels, which are processed from left to right.
- each pixel in the current image block is quantized in turn, and the quantization error is diffused and transmitted to the pixels in the current image block and outside the current image block that have not been quantized. .
- Figure 6 (a) shows the comparison of the three methods in PSNR performance
- Figure 6 (b) shows the comparison of the three methods in MSSIM performance.
- the main difference is the probability visual password or random grid visual password in Figure 2.
- quantization and diffusion are also performed on pixels, and in FIG. 3, quantization is specifically vector quantization, and diffusion is also vector diffusion, that is, diffusion into surrounding image blocks.
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Facsimile Transmission Control (AREA)
- Two-Way Televisions, Distribution Of Moving Picture Or The Like (AREA)
- Image Processing (AREA)
Abstract
一种灰度图像可视加密方法,该方法将可视密码中产生分存图像和叠加解码两个过程置于一个误差扩散反馈环内。在编码器端,按照一定扫描顺序来顺序处理每个像素或者图像块,对产生的分存图像上的每个像素或者图像块,即时检查其叠加解码效果,将叠加后获得的重建目标图像和原始秘密图像对比,获得二者的差异,将对比差异反馈并扩散到尚未量化的周围像素或者图像块上,从而达到提高解码重建秘密图像质量的目的,分析过程对应于从秘密图像产生分存图像过程,而合成过程对应于从分存图像获得目标图像过程,该方法将合成放在分析中,即基于合成的分析。
Description
本发明涉及多媒体通信安全领域,尤其涉及一种灰度图像可视加密方法。
可视加密是一种解码过程高效且不需要任何计算的秘密分享方法,对于一个(n,k)阈值可视密码,它将一幅秘密图像分成n份传输,只有获得其中的至少k份,将其叠加才能恢复秘密图像的内容。近年来,视觉密码的研究有两个重点:一个是减小分存图像的尺寸,另一个是提高重建图像的视觉质量。其中,分存图像尺寸相对于秘密图像尺寸一般有扩展,小的扩展系数甚至尺寸不扩展的可视密码具备很多优势,例如处理复杂度低,占用存储空间小,占用传输带宽低等。针对二值秘密图像,典型的尺寸不变视觉密码算法包括随机网格法、概率算法和分块编码方法等三种。但是,在减小分存图像尺寸的同时,重建图像的质量却大幅下降。尤其对于灰度秘密图像,其重建图像的质量远达不到要求。
与普通二值图像视觉密码不同,灰度和彩色图像视觉密码的秘密图像是灰度和彩色图像,重建图像是呈现类似半色调的图像。聚焦到灰度图像视觉密码,当前对尺寸不扩展视觉密码(Size-Invariant Visual Cryptography)的研究可以大致划分为两类。一类是半色调化和视觉密码加密分离的方案,另一类是直接分块量化和映射的方案。其中:
第一类关于半色调化和视觉密码加密分离的方案:
首先将灰度图像转换为二值图像,例如使用半色调化或者简单的固定阈值二值量化技术,然后对此二值图像进行视觉密码编码。该方案将半色调过程和视觉加密过程分隔开:半色调化过程不考虑后续的视觉密码加密;而视觉加密的时候不考虑所处理的图像是否是半色调的,还是类似公司标识那样的二值图像。这类算法是目前灰度图像视觉密码的主流。
第二类关于直接分块量化和映射的方案:
首先将灰度图像分块,然后将每块的平均灰度量化为不同的量化级,这些量化级对应着重建图像上不同的二值模式,每个二值模式具有不同的块平均灰度,从而重建图像可通过不同二值模式来近似表达秘密图像的局部亮度和色彩。
现有的框架存在两个主要问题:
1.不补偿重建图像上的亮度损失,2.不直接优化重建图像质量。
首先,从秘密像素/图像块到重建像素/图像块之间的有损映射没有被周围其他的像素和图像块补偿。由于视觉密码的安全性条件的限制,重建图像上表达秘密二值图像的能力有限。 例如,如果采用2乘2的图像块,根据块内黑像素的数量,秘密图像的图像块有5种,但是,在重建图像上,只允许有3种可能的图像块,其中黑像素的数量分别是2、3和4。这样,从秘密图像块到重建图像块,存在多对一的有损映射。目前的算法都无法补偿这种亮度损失。
其次,现有算法没有直接优化重建图像的感知质量。不仅如此,虽然重建图像是半色调类型的图像,但是,现有算法所用的图像质量测度并没有反映人的视觉系统如何感知半色调图像。具体而言,重建图像的谱特性没有反映在现有的质量测度中。
现有文献中常用的质量测度是块内方差和全局对比度。块内方差测度度量了重建图像相对于秘密二值图像的均匀性。但是,它并没有度量半色调重建图像的感知质量。全局对比度测度则度量了重建图像上用于表达黑色秘密像素和白色秘密像素的图像块的亮度差异,该测度是为二值秘密图像设计的。近期,X.T.Wu等人提出能产生蓝噪声的分存图像的视觉密码,但是,将两个蓝噪声分存图像叠加后,重建图像将不再具有蓝噪声特性,所以该类算法只适合将灰度秘密图像简单二值化后再加密,从而具有很低的细节分辨能力。
发明内容
本发明的目的在于提出一种灰度图像可视加密方法,以提高重建的目标图像的视觉质量。
本发明为了实现上述目的,采用如下技术方案:
一种灰度图像可视加密方法,包括如下步骤:
s11.色域映射
输入秘密图像是灰度图像J[n],其中n=[n
x,n
y],n
x是像素的水平位置,n
x是像素的垂直位置,确定图像块的大小为B×B;
将灰度图像J[n]的色域映射到可视密码叠加重建图像(即目标图像)的色域中;
s12.分块
将色域映射后的图像进行分块,分割为多个互相不重叠、大小为B×B的图像块;
以光栅扫描的顺序依次访问各个图像块,每次处理一个图像块;
s13.矢量量化
对每个图像块,使用矢量量化器量化为二值图像块,即量化后每个像素取值为0或者1;
s14.矢量可视加密
将量化后的二值图像块,进行矢量可视密码编码,获得k个分存图像块s
1[n],…,s
k[n];
s15.模拟叠加解码
将k个分存图像块s
1[n],…,s
k[n]进行叠加,即:
s16.误差计算
s17.矢量误差扩散
使用矢量误差扩散滤波器H[n]将误差e[n]扩散到当前图像块周围的四个图像块中;
s18.利用上述步骤s13-s17处理每个图像块,当处理完所有的图像块之后,将k个分存图像s
1[n],…,s
k[n]作为输出,获得了可视加密方法的输出结果。
优选地,所述步骤s11具体为:
s111.计算图像块内像素数:m=B
2;
优选地,所述步骤s13具体为:
s132.每个像素均使用二值标量量化器量化,量化阈值设为1/2,输出重建点设为0和1;
s133.将二值标量量化器的输入输出误差利用误差扩散滤波器进行扩散,扩散到该图像块中还没有量化的其他像素上;
s134.处理完所有像素,输出量化后的图像块,作为矢量量化器的输出。
优选地,所述步骤s14具体为:
s141.使用任何一种(k,k)阈值可视密码的基本矩阵A
0和A
1来构建基本矩阵C
j,其中:
D是目标图像块中所有可能的白像素数目的集合,|D|表示集合D中元素数目;
s142.根据输入秘密图像的图像块中白像素的数量,选择矩阵C
j;
将矩阵C
j的列置乱,将其每一行重新排列为B×B图像块,分配给一个分存图像,作为这个分存图像的对应图像块,这样就获得了k个分存图像块s
1[n],…,s
k[n]。
优选地,所述步骤s142中,矩阵C
j的选择方法为:
对于一个秘密图像的图像块,图像块内像素数是m,则所有可能白像素数集合是M={0,1,…,m};而步骤s141中得到的基本矩阵的集合是Γ={C
j:j=0,1,…,|D-1|};
据此,设计一个从M到Γ的映射Φ:M→Γ,只需要满足:
根据这个映射,给定秘密图像块内白像素数量,选取一个矩阵C
j。
优选地,所述步骤s17具体为:
s171.邻域选择
选择当前图像块的右边、左下、下面、右下四个图像块为当前图像块的邻域;这些邻域是当前图像块误差扩散的目的地;
s172.将误差e[n]使用矢量误差扩散滤波器H[n]扩散到邻域中的图像块中:
顺序访问当前图像块中的每个像素,计算重建图像像素和矢量量化器输入像素的误差,使用普通标量误差扩散滤波器将该误差扩散到当前图像块周围的像素上;
如果扩散到一个已经量化的像素上,则不改变该像素值,该像素将扩散来的误差原封不动,使用同样的标量误差扩散滤波器传递到它周围的像素上。
优选地,所述方法适用于以下三类针对二值图像设计的可视密码的任一种,即:
概率可视密码、随机格点可视密码和矢量可视密码。
本发明具有如下优点:
(1)本发明方法可以方便地和现有的针对二值图像的概率可视密码、随机格点可视密码以及矢量可视密码中的任一种结合,构建针对灰度图像的可视密码,并可显著提高重建的目标图像的视觉质量。(2)本发明方法满足灰度可视密码的对比度和安全性条件,只要基本的二值可视密码是完善安全的,那么该方法获得的灰度图像可视密码就是完善安全的。
图1为本发明实施例中灰度图像可视加密方法的原理框图。
图2为本发明实施例中灰度图像可视加密方法的一种流程框图。
图3为本发明实施例中灰度图像可视加密方法的另一种流程框图。
图4为本发明实施例中使用标量量化和块内误差扩散实现矢量量化示意图。
图5为本发明实施例中矢量误差扩散示意图。
图6为本发明实施例中方法与Wang,Yang等人算法对比结果图。
本发明的基本思想如下:
如图1所示,提出一个具有闭环反馈的分析-合成(Analysis-by-Synthesis,AbS)框架的加密器,从分析-合成的角度设计视觉密码,即将秘密图像分割为分存图像的过程是其中的分析部分;而将分存图像在加密器叠加生成重建图像是其中的合成部分。
AbS的基本含义是在加密器中(而不是在解码器中),对加密获得的分存图像解码,将解码结果作为优化视觉质量的依据。具体而言,在这个AbS框架中,将加密器合成的重建图像反馈到分析部分,获得当前重建像素(或图像块)与秘密像素(或图像块)之间差异,这样,半色调化过程可以仔细调整其产生的二值图像中的黑白像素的比例,从而补偿上述差异。从结构上看,将可视加密和解码移到了误差扩散半色调算法的反馈环之内。
本发明提出的AbS框架具有很高的灵活性,它可以和现有的三种尺寸不变可视密码结合,即可以和概率可视密码、随机格点可视密码以及矢量可视密码中的任一种结合,分别获得基于AbS的概率可视密码、基于AbS的随机格点可视密码和基于AbS的矢量可视密码。所获得的加密系统具有香农所定义的完善安全性(Perfect Secrecy)。
下面结合附图以及具体实施方式对本发明作进一步详细说明:
如图3所示,以本发明方法与矢量可视密码结合为例,说明灰度图像可视加密流程:
一种灰度图像可视加密方法,包括如下步骤:
s11.色域映射
输入秘密图像是灰度图像J[n],其中n=[n
x,n
y],n
x是像素的水平位置,n
x是像素的垂直位置,确定图像块大小为B×B;
将灰度图像J[n]的色域映射到可视密码叠加重建图像(即目标图像)的色域中。
该步骤s11具体为:
s111.计算图像块内像素数:m=B
2。
使用上述步骤的色域映射,能够保证整个误差扩散反馈环的稳定性。
如果需要的话(例如可以是当输入秘密图像的对比度较低时),可先对输入图像进行直方 图均衡,再使用上述色域映射。
s12.分块
将色域映射后的图像进行分块,分割为多个互相不重叠、大小为B×B的图像块。
以光栅扫描的顺序依次访问各个图像块,每次处理一个图像块。
s13.矢量量化
对每个图像块,使用矢量量化器量化为二值图像块,即量化后每个像素取值为0或者1。
该步骤s13具体为:
s132.对每个像素均使用二值量化器量化,量化阈值设为1/2,输出重建点设为0和1。
s133.将二值标量量化器的输入输出误差利用误差扩散滤波器进行扩散,扩散到该图像块中还没有量化的其他像素上。
s134.处理完所有像素,输出量化后的图像块,作为矢量量化器的输出。
使用矢量量化器,能够以复杂度较低的标量量化和误差扩散实现复杂度较高的矢量量化。
s14.矢量可视加密
将量化后的二值图像块,进行矢量可视密码编码,获得k个分存图像块s
1[n],…,s
k[n]。
该步骤s14具体为:
s141.使用任何一种(k,k)阈值可视密码的基本矩阵A
0和A
1来构建基本矩阵C
j,其中:
D是目标图像块中所有可能的白像素数目的集合,|D|表示集合D中元素数目。
s142.根据输入秘密图像的图像块中白像素的数量,选择矩阵C
j。
矩阵C
j的选择方法为:
对于一个秘密图像的图像块,图像块内像素数是m,则所有可能白像素数集合是M={0,1,…,m};而步骤s141中得到的基本矩阵的集合是Γ={C
j:j=0,1,…,|D-1|}。
据此,设计一个从M到Γ的映射Φ:M→Γ,只需要满足:
根据这个映射,给定秘密图像块内白像素数量,选取一个矩阵C
j。
例如,当M={0,1,…,4},Γ={C
j:j=0,1,2},可以使用表1的映射,即满足上述条件。
表1
m | Φ(m) |
0 | 0 |
1 | 0 |
2 | 1 |
3 | 2 |
4 | 2 |
使用这个映射,能够保证白像素多的秘密图像块对应于白像素多的目标图像块。
将矩阵C
j的列置乱,将其每一行重新排列为B×B图像块,分配给一个分存图像,作为这个分存图像的对应图像块,这样就获得了k个分存图像块s
1[n],…,s
k[n]。
使用该构建方案,能够根据目标图像中白像素数目,设计出符合要求的基本矩阵。
s15.模拟叠加解码
将k个分存图像块s
1[n],…,s
k[n]进行叠加,即:
s16.误差计算
s17.矢量误差扩散
使用矢量误差扩散滤波器H[n]将误差e[n]扩散到当前图像块周围的四个图像块中;
该步骤s17具体为:
s171.邻域选择
选择当前图像块的右边、左下、下面、右下四个图像块为当前图像块的邻域,这些邻域是当前图像块误差扩散的目的地。
s172.将误差e[n]使用矢量误差扩散滤波器H[n]扩散到邻域中的图像块中:
顺序访问当前图像块中的每个像素,计算重建图像像素和矢量量化器输入像素的误差,使用普通标量误差扩散滤波器将该误差扩散到当前图像块周围的像素上;
该普通标量误差扩散滤波器例如采用常用的Floyd-Steinberg滤波器。
如果扩散到一个已经量化的像素上,则不改变该像素值,该像素将扩散来的误差原封不动,使用同样的标量误差扩散滤波器传递到它周围的像素上。
使用该方案,能够用一系列复杂度低的标量误差扩散实现高维的矢量误差扩散。
s18.利用上述步骤s13-s17处理每个图像块,当处理完所有的图像块之后,将分存图像s
1[n],…,s
k[n]作为输出,获得了可视加密方法的输出结果。
通过上述流程,使得本发明方法与矢量可视密码有机结合,构建针对灰度图像的可视密码,并可显著提高重建的目标图像的视觉质量。
在进行矢量量化的过程中,使用标量量化器结合块内误差扩散来实现。这个实现的示意图如图4所示。斜线框内表示已量化的像素,白色框内表示尚未量化,从左往右依次处理。
针对矢量误差扩散,可使用序贯标量量化和误差扩散实现,如图5所示。
从图5(a)到图5(d)依次量化当前图像块(斜线框)内的每个像素,并将量化误差扩散和传递到当前图像块内和当前图像块外尚未量化的像素上。
使用基本矩阵构建方法,针对k=2,m=4,获得的基本矩阵如下:
类似的,对于k=3,m=4,可获得如下的基本矩阵:
目标图像的质量,发明人与现有的两种算法做了比较,分别是:
1)D.S.Wang,F.Yi,and X.B.Li,“Probabilistic visual secret sharing schemes for grey-scale images and color images,”Inform.Sci.,vol.181,no.11,pp.2189–2208,Jun.2011.
2)C.N.Yang,“New visual secret sharing schemes using probabilistic method,”Pattern Recognition Lett.,vol.25,no.4,pp.481–494,Mar.2004.
与Wang,Yang等人算法对比,本发明目标图像的质量具有显著性能提升,如图6所示。图6(a)为三种方法在PSNR性能上的对比,图6(b)为三种方法在MSSIM性能上的对比。
在对比时使用了Kodak数据库中24幅自然图像,横坐标Image index是24幅图像的索引。
从图6可以看出,本发明方法无论是HPSNR(Human Peak Signal to Noise Ratio,峰值信噪比)还是MSSIM(Mean Structural Similarity Measure,结构相似度)都有显著提升。
如图2所示,本发明实施例还针对概率可视密码或者随机网格可视密码,给出了分析- 合成框架的具体实施例子,其中,图像块的大小是B=1。
从图2与图3对比可以看出,针对概率可视密码或者随机网格可视密码的灰度图像可视加密方法的流程与针对矢量可视密码的灰度图像可视加密方法的流程是一致的。
主要区别在于图2中的概率可视密码或者随机网格可视密码,其图像块的大小是B=1,这样每个图像块就变成了一个像素块。由此,可视加密和解密部分都是针对每个像素进行的,而不是像图3那样针对一个图像块进行。另外,在图2中,量化和扩散也是针对像素进行的,而在图3中,量化具体是矢量量化,扩散也是矢量扩散,即扩散到周围的图像块中。
当然,以上说明仅仅为本发明的较佳实施例,本发明并不限于列举上述实施例,应当说明的是,任何熟悉本领域的技术人员在本说明书的教导下,所做出的所有等同替代、明显变形形式,均落在本说明书的实质范围之内,理应受到本发明的保护。
Claims (7)
- 一种灰度图像可视加密方法,其特征在于,包括如下步骤:s11.色域映射输入秘密图像是灰度图像J[n],其中n=[n x,n y],n x是像素的水平位置,n x是像素的垂直位置,确定图像块的大小为B×B;将灰度图像J[n]的色域映射到可视密码叠加重建图像(即目标图像)的色域中;s12.分块将色域映射后的图像进行分块,分割为多个互相不重叠、大小为B×B的图像块;以光栅扫描的顺序依次访问各个图像块,每次处理一个图像块;s13.矢量量化对每个图像块,使用矢量量化器量化为二值图像块,即量化后每个像素取值为0或者1;s14.矢量可视加密将量化后的二值图像块,进行矢量可视密码编码,获得k个分存图像块s 1[n],…,s k[n];s15.模拟叠加解码将k个分存图像块s 1[n],…,s k[n]进行叠加,即:s16.误差计算s17.矢量误差扩散使用矢量误差扩散滤波器H[n]将误差e[n]扩散到当前图像块周围的四个图像块中;s18.利用上述步骤s13-s17处理每个图像块,当处理完所有的图像块之后,将k个分存图像s 1[n],…,s k[n]作为输出,获得了可视加密方法的输出结果。
- 根据权利要求1所述的灰度图像可视加密方法,其特征在于,所述步骤s17具体为:s171.邻域选择选择当前图像块的右边、左下、下面、右下四个图像块为当前图像块的邻域;这些邻域是当前图像块误差扩散的目的地;s172.将误差e[n]使用矢量误差扩散滤波器H[n]扩散到邻域中的图像块中:顺序访问当前图像块中的每个像素,计算重建图像像素和矢量量化器输入像素的误差,使用普通标量误差扩散滤波器将该误差扩散到当前图像块周围的像素上;如果扩散到一个已经量化的像素上,则不改变该像素值,该像素将扩散来的误差原封不动,使用同样的标量误差扩散滤波器传递到它周围的像素上。
- 根据权利要求1至6任一项所述的灰度图像可视加密方法,其特征在于,所述方法适用于以下三类针对二值图像设计的可视密码的任一种,即:概率可视密码、随机格点可视密码和矢量可视密码。
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
US16/475,545 US11115194B2 (en) | 2018-08-08 | 2018-08-24 | Gray image visual encryption method |
Applications Claiming Priority (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810894261.0 | 2018-08-08 | ||
CN201810894261.0A CN109214971B (zh) | 2018-08-08 | 2018-08-08 | 一种灰度图像可视加密方法 |
Publications (1)
Publication Number | Publication Date |
---|---|
WO2020029334A1 true WO2020029334A1 (zh) | 2020-02-13 |
Family
ID=64988178
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
PCT/CN2018/102211 WO2020029334A1 (zh) | 2018-08-08 | 2018-08-24 | 一种灰度图像可视加密方法 |
Country Status (3)
Country | Link |
---|---|
US (1) | US11115194B2 (zh) |
CN (1) | CN109214971B (zh) |
WO (1) | WO2020029334A1 (zh) |
Cited By (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN112422767A (zh) * | 2020-10-30 | 2021-02-26 | 国家电网有限公司大数据中心 | 一种基于视觉效果秘密共享方法及系统 |
US11115194B2 (en) | 2018-08-08 | 2021-09-07 | Shandong University Of Science And Technology | Gray image visual encryption method |
CN115150818A (zh) * | 2022-09-05 | 2022-10-04 | 光谷技术有限公司 | 一种基于人工智能的通信传输加密方法 |
Families Citing this family (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111756770B (zh) * | 2020-07-16 | 2022-05-31 | 鸣飞伟业技术有限公司 | 一种高清视频安全隔离显示网关 |
CN112084513B (zh) * | 2020-08-28 | 2022-03-04 | 山东科技大学 | 一种彩色图像的可视加密方法 |
CN113222829B (zh) * | 2021-02-25 | 2023-04-25 | 安徽师范大学 | 基于Bernstein基的数字图像分存方法及图像复原方法 |
CN115733933B (zh) * | 2022-11-29 | 2023-10-31 | 南京先维信息技术有限公司 | 基于人工智能的网络安全加密方法 |
CN118314221B (zh) * | 2024-06-07 | 2024-08-20 | 安徽全采智能科技有限公司 | 一种道路交通的雷视一体机标定及拼接方法 |
CN118474268A (zh) * | 2024-07-10 | 2024-08-09 | 广东工业大学 | 一种基于物理掺杂氮杂菲结构的光学图像加密、解密方法 |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101751658A (zh) * | 2009-12-09 | 2010-06-23 | 清华大学 | 具有最小像素膨胀和最优对照度的图像可视分存方法 |
US20110194727A1 (en) * | 2010-02-11 | 2011-08-11 | National Taiwan University Of Science & Technology | Image data processig systems for hiding secret information and data hiding methods using the same |
CN104463767A (zh) * | 2014-11-27 | 2015-03-25 | 山东科技大学 | 一种具有伪装图像和篡改认证功能的可视加密及解密方法 |
CN104574262A (zh) * | 2015-01-16 | 2015-04-29 | 山东科技大学 | 一种扩展可视加密和解密方法 |
CN105404817A (zh) * | 2015-10-27 | 2016-03-16 | 西安电子科技大学 | 基于(k, n)门限的用户友好的可视秘密共享方法 |
Family Cites Families (28)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
AU2587392A (en) * | 1991-11-05 | 1993-06-07 | Virtual Image Group, L.P. | Optical image encryption and decryption processes |
JP2002057882A (ja) * | 2000-04-21 | 2002-02-22 | Sony Corp | 情報埋め込み装置及び情報埋め込み方法、情報処理装置及び情報処理方法、コンテンツ処理装置及びコンテンツ処理方法、監視装置及び監視方法、並びに、記憶媒体 |
US6802613B2 (en) * | 2002-10-16 | 2004-10-12 | Eastman Kodak Company | Broad gamut color display apparatus using an electromechanical grating device |
JP2006510051A (ja) * | 2002-12-16 | 2006-03-23 | コーニンクレッカ フィリップス エレクトロニクス エヌ ヴィ | 光の偏光を用いる視覚暗号化を有する認証システム |
CN1759416A (zh) * | 2003-03-11 | 2006-04-12 | 皇家飞利浦电子股份有限公司 | 视觉密码术系统 |
KR20060040698A (ko) * | 2003-07-21 | 2006-05-10 | 코닌클리케 필립스 일렉트로닉스 엔.브이. | 영상 정렬 |
GB0323364D0 (en) * | 2003-10-07 | 2003-11-05 | Koninkl Philips Electronics Nv | Image decryption system |
EP1884059A2 (en) * | 2005-05-13 | 2008-02-06 | Temple University of the Commonwealth System of Higher Education | Secret sharing technique with low overhead information content |
US7408558B2 (en) * | 2005-08-25 | 2008-08-05 | Eastman Kodak Company | Laser-based display having expanded image color |
GB0522968D0 (en) * | 2005-11-11 | 2005-12-21 | Popovich Milan M | Holographic illumination device |
JP2008131282A (ja) * | 2006-11-20 | 2008-06-05 | Sony Corp | 映像伝送方法、映像伝送システム及び映像処理装置 |
JP5176655B2 (ja) * | 2008-03-31 | 2013-04-03 | 富士通株式会社 | 画像復号化装置 |
EP2120448A1 (en) * | 2008-05-14 | 2009-11-18 | Thomson Licensing | Method of processing of a compressed image into a gamut mapped image using spatial frequency analysis |
FR2932046B1 (fr) * | 2008-06-03 | 2010-08-20 | Thales Sa | Procede et systeme permettant de crypter visuellement les objets mobiles au sein d'un flux video compresse |
FR2932045B1 (fr) * | 2008-06-03 | 2010-08-20 | Thales Sa | Procede et systeme permettant de proteger des la compression la confidentialite des donnees d'un flux video lors de sa transmission |
CN101447071B (zh) * | 2009-01-04 | 2010-09-15 | 清华大学 | 图案的视觉加密防伪方法 |
WO2011111191A1 (ja) * | 2010-03-10 | 2011-09-15 | 富士通株式会社 | 画像復号装置、画像復号方法及び画像復号用コンピュータプログラム |
US8966656B2 (en) * | 2011-10-21 | 2015-02-24 | Blackberry Limited | Displaying private information using alternate frame sequencing |
US20130135588A1 (en) * | 2011-11-29 | 2013-05-30 | Milan Momcilo Popovich | 3D display apparatus |
US20150070402A1 (en) * | 2013-09-12 | 2015-03-12 | Qualcomm Incorporated | Real-time color calibration of displays |
US20150106950A1 (en) * | 2013-10-10 | 2015-04-16 | Elwha Llc | Methods, systems, and devices for handling image capture devices and captured images |
US9838388B2 (en) * | 2014-08-26 | 2017-12-05 | Veridium Ip Limited | System and method for biometric protocol standards |
US20170061851A1 (en) * | 2014-04-30 | 2017-03-02 | Hewlett-Packard Development Company, L.P. | Large gamut pixel and subtractive mask for a visual presentation |
CN104281993A (zh) * | 2014-07-29 | 2015-01-14 | 山东科技大学 | 基于可视加密和特征点匹配的抗旋转攻击数字水印方法 |
US20160224528A1 (en) * | 2015-01-30 | 2016-08-04 | Technology Happens LLC | Method and System for Collaborative, Streaming Document Sharing with Verified, On-Demand, Freestyle Signature Process |
US20180174528A1 (en) * | 2016-12-19 | 2018-06-21 | Amazon Technologies, Inc. | Control system for an electrowetting display device with rendering engine |
JP6669994B2 (ja) * | 2017-03-16 | 2020-03-18 | 京セラドキュメントソリューションズ株式会社 | 画像形成装置、画像形成方法及び画像形成プログラム |
CN109214971B (zh) | 2018-08-08 | 2019-05-28 | 山东科技大学 | 一种灰度图像可视加密方法 |
-
2018
- 2018-08-08 CN CN201810894261.0A patent/CN109214971B/zh not_active Expired - Fee Related
- 2018-08-24 WO PCT/CN2018/102211 patent/WO2020029334A1/zh active Application Filing
- 2018-08-24 US US16/475,545 patent/US11115194B2/en active Active
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101751658A (zh) * | 2009-12-09 | 2010-06-23 | 清华大学 | 具有最小像素膨胀和最优对照度的图像可视分存方法 |
US20110194727A1 (en) * | 2010-02-11 | 2011-08-11 | National Taiwan University Of Science & Technology | Image data processig systems for hiding secret information and data hiding methods using the same |
CN104463767A (zh) * | 2014-11-27 | 2015-03-25 | 山东科技大学 | 一种具有伪装图像和篡改认证功能的可视加密及解密方法 |
CN104574262A (zh) * | 2015-01-16 | 2015-04-29 | 山东科技大学 | 一种扩展可视加密和解密方法 |
CN105404817A (zh) * | 2015-10-27 | 2016-03-16 | 西安电子科技大学 | 基于(k, n)门限的用户友好的可视秘密共享方法 |
Cited By (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11115194B2 (en) | 2018-08-08 | 2021-09-07 | Shandong University Of Science And Technology | Gray image visual encryption method |
CN112422767A (zh) * | 2020-10-30 | 2021-02-26 | 国家电网有限公司大数据中心 | 一种基于视觉效果秘密共享方法及系统 |
CN112422767B (zh) * | 2020-10-30 | 2022-07-22 | 国家电网有限公司大数据中心 | 一种基于视觉效果秘密共享方法及系统 |
CN115150818A (zh) * | 2022-09-05 | 2022-10-04 | 光谷技术有限公司 | 一种基于人工智能的通信传输加密方法 |
CN115150818B (zh) * | 2022-09-05 | 2022-11-04 | 光谷技术有限公司 | 一种基于人工智能的通信传输加密方法 |
Also Published As
Publication number | Publication date |
---|---|
US20210044431A1 (en) | 2021-02-11 |
US11115194B2 (en) | 2021-09-07 |
CN109214971A (zh) | 2019-01-15 |
CN109214971B (zh) | 2019-05-28 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
WO2020029334A1 (zh) | 一种灰度图像可视加密方法 | |
Yan et al. | Improving the visual quality of size-invariant visual cryptography for grayscale images: an analysis-by-synthesis (AbS) approach | |
Askari et al. | An extended visual cryptography scheme without pixel expansion for halftone images | |
Ozdemir et al. | Fuzzy algorithms for combined quantization and dithering | |
JP2010211214A (ja) | カラー画像における色量子化方法 | |
Alex et al. | Enhanced image secret sharing via error diffusion in halftone visual cryptography | |
EP3349446A1 (fr) | Procédé et dispositif d'encodage dynamique contrôlé d'un signal numérique multidimensionnel, en particulier un signal d'image et procédé et dispositif correspondant de décodage | |
Chan et al. | A framework of reversible color-to-grayscale conversion with watermarking feature | |
JP2007065916A (ja) | 画像処理方法、画像処理装置、画像形成装置、コンピュータプログラム及び記録媒体 | |
Fan et al. | Flexible patch moving modes for pixel-value-ordering based reversible data hiding methods | |
US9990686B2 (en) | Method for generating a data-bearing halftone image, and method for decoding the data-bearing halftone image | |
Thomas et al. | Halftone visual cryptography for grayscale images using error diffusion and direct binary search | |
JPH04144485A (ja) | 画像処理装置 | |
JP2007306513A (ja) | 画像データの圧縮方法および装置 | |
Tamilarasi et al. | Improving image quality in extended visual cryptography for halftone images with no pixel expansion | |
Wu et al. | Threshold visual cryptography scheme for color images with no pixel expansion | |
Wang et al. | Halftone visual cryptography by iterative halftoning | |
Yamaguchi | An extended visual cryptography scheme for continuous-tone images | |
Hong et al. | A difference matching technique for data embedment based on absolute moment block truncation coding | |
Hu | Synergistic compensation for RGB-based blind color image watermarking to withstand JPEG compression | |
Felix et al. | Self-embedding reversible color-to-grayscale conversion with watermarking feature | |
JP4027300B2 (ja) | 画像処理方法、画像処理装置、画像形成装置、コンピュータプログラム及び記録媒体 | |
CN111953485B (zh) | 一种门限可变的秘密图像共享方法 | |
Zhou et al. | Block Halftoning for Size-Invariant Visual Cryptography Based on Two-Dimensional Lattices | |
JP4549306B2 (ja) | 画像処理方法、画像処理装置、画像形成装置、及びコンピュータプログラム |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
121 | Ep: the epo has been informed by wipo that ep was designated in this application |
Ref document number: 18929181 Country of ref document: EP Kind code of ref document: A1 |
|
NENP | Non-entry into the national phase |
Ref country code: DE |
|
122 | Ep: pct application non-entry in european phase |
Ref document number: 18929181 Country of ref document: EP Kind code of ref document: A1 |