CN108876691B - Self-adaptive reversible watermarking method and system - Google Patents
Self-adaptive reversible watermarking method and system Download PDFInfo
- Publication number
- CN108876691B CN108876691B CN201710326906.6A CN201710326906A CN108876691B CN 108876691 B CN108876691 B CN 108876691B CN 201710326906 A CN201710326906 A CN 201710326906A CN 108876691 B CN108876691 B CN 108876691B
- Authority
- CN
- China
- Prior art keywords
- pixel
- pair
- optimal path
- difference value
- subset
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Classifications
-
- 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
- G06T2201/00—General purpose image data processing
- G06T2201/005—Image watermarking
- G06T2201/0203—Image watermarking whereby the image with embedded watermark is reverted to the original condition before embedding, e.g. lossless, distortion-free or invertible watermarking
Abstract
The invention provides a self-adaptive reversible watermarking method and a system, wherein the method comprises the following steps: dividing the carrier image into a plurality of non-overlapping adjacent pixel pairs; calculating corresponding difference value pairs by using a gradient prediction method and generating a two-dimensional difference value histogram; selecting an optimal path in the two-dimensional difference histogram by adopting a dynamic programming algorithm, and establishing a first corresponding relation between a pixel pair corresponding to a difference value pair on the optimal path and bit information carried by a watermark to be embedded; if the difference value pair corresponding to the pixel pair is not on the optimal path, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is on the optimal path, modifying the pixel pair according to a preset first rule if the value of the bit information corresponding to the pixel pair is 1 according to the first corresponding relation; and if the value of the bit information corresponding to the pixel pair is 0, the pixel pair is not modified. The scheme has high data embedding capacity, low embedding distortion and high safety of hidden data.
Description
Technical Field
The invention relates to the technical field of digital watermarking, in particular to a self-adaptive reversible watermarking method and a self-adaptive reversible watermarking system.
Background
With the modern society entering the internet information era, multimedia digital carriers such as images, audio, video and the like become main media for bearing information. The rapid development of computer science and technology and the popularity of the internet have enabled more and more information to be generated, distributed and shared widely by these multimedia carriers. Accordingly, the openness and the shareability of the internet era cause a series of security problems, such as illegal copying, illegal tampering and the like, to digital information in the process of transmission. In addition, the hiding and protection of confidential information that the propagator does not want to disclose is also an important requirement. In response to such problems, information hiding techniques have been developed.
Research related to information hiding technology is widely concerned by scholars in recent decades, but most of the technologies such as steganography and robust watermarking technology damage the original carrier while embedding secret information, so that the embedded carrier cannot be restored without loss after extracting the secret information. However, in some sensitive scenarios, such as medical image processing, it is also important to recover the original medical image without loss while obtaining patient privacy information. Therefore, the reversible digital watermarking technology is the subject of the next intense research. Since there is a large amount of redundant information in a natural image, even if some pixels are modified due to embedded information, they can still be recovered using the relevant pixel information, which is why this technique has a reversible nature. The criteria for measuring the quality of the reversible watermarking algorithm have two aspects: embedding distortion and embedding capacity. By comparing the embedding carrier with the original carrier, the lower the embedding distortion, the better the concealment of the embedding algorithm, i.e. the more similar the embedding carrier is to the original carrier. The embedding capacity indicates the number of information bits embedded into the carrier.
The existing reversible watermarking method based on the histogram mainly constructs a statistical histogram through pixel points of a carrier image, divides the pixel points in the image into an embeddable point set in which watermark data are embedded and a translational movable point set according to the statistical histogram, executes watermark embedding processing on all the pixel points in the embeddable point set, and executes translation processing on all the pixel points in the movable point set. The distribution of the pixel histogram of the carrier image is uniform, so that the data embedding capacity of the method is limited, the embedding distortion is high, and the safety of hidden data is further reduced.
Disclosure of Invention
The invention provides a self-adaptive reversible watermarking method and a self-adaptive reversible watermarking system, which are used for solving the problems of low data embedding capacity, high embedding distortion and low security of hidden data in the watermarking method in the prior art.
A first aspect of the present invention provides an adaptive reversible watermarking method, comprising:
dividing the carrier image into a plurality of pixel pairs (x, y) which are not overlapped with each other, wherein two pixel points in each pixel pair are adjacent, wherein x is the pixel value of the previous pixel point, and y is the pixel value of the next pixel point;
for each pixel pair (x, y), calculating a corresponding prediction value z according to the neighborhood of the pixel pair by adopting a gradient prediction method, and calculating a corresponding difference value pair (d) by utilizing a first formulax,dy) The first formula is:
d1＝x-z，d2＝y-z；
generating a two-dimensional difference value histogram according to the difference value pairs corresponding to the plurality of pixel pairs;
selecting a starting point and an end point in the two-dimensional difference histogram, and selecting an optimal path in a grid formed by the starting point and the end point by adopting a dynamic programming algorithm, wherein the optimal path is a path with the maximum sum of the frequencies of first difference values passing through all paths from the starting point to the end point in the grid, the actual embeddable capacity of a pixel pair corresponding to the first difference value pair passing through the optimal path is not less than the capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in the horizontal direction or the vertical direction;
defining the first difference value pair as an expansion difference value pair, and defining the rest difference value pairs as translation difference value pairs;
establishing a first corresponding relation between the pixel pair corresponding to the expansion difference value pair and bit information carried by the watermark to be embedded;
for each pixel pair, if the difference value pair corresponding to the pixel pair is a translation difference value pair, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; and if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair.
A second aspect of the invention provides an adaptive reversible watermarking system comprising:
the image processing device comprises a pixel pair module, a pixel value acquisition module and a pixel value acquisition module, wherein the pixel pair module is used for dividing a carrier image into a plurality of pixel pairs (x, y) which are not overlapped with each other, and two pixel points in each pixel pair are adjacent, wherein x is the pixel value of a previous pixel point, and y is the pixel value of a next pixel point;
a difference value pair module, for calculating a corresponding prediction value z according to the neighborhood of the pixel pair by using a gradient prediction method and calculating a corresponding difference value pair (d) by using a first formula for each pixel pair (x, y)x,dy) The first formula is:
d1＝x-z，d2＝y-z；
the histogram module is used for generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs;
an optimal path module, configured to select a starting point and an end point in the two-dimensional difference histogram, and select an optimal path in a grid formed by the starting point and the end point by using a dynamic programming algorithm, where the optimal path is a path in which a sum of frequencies of first difference values passing through all paths from the starting point to the end point in the grid is the largest, an actual embeddable capacity of a pixel pair corresponding to the first difference value passing through the optimal path is not less than a capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in a horizontal direction or a vertical direction;
a defining module, configured to define the first difference pair as an expansion difference pair, and define the remaining difference pairs as translation difference pairs;
a corresponding module, configured to establish a first corresponding relationship between the pixel pair corresponding to the expanded difference value pair and bit information carried by the watermark to be embedded;
the modification module is used for modifying the pixel pair according to a preset first rule if the difference value pair corresponding to the pixel pair is a translation difference value pair aiming at each pixel pair; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; and if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair.
The self-adaptive reversible watermarking method and the self-adaptive reversible watermarking system have the advantages that a carrier image is divided into a plurality of non-coincident adjacent pixel pairs; calculating corresponding difference value pairs by using a gradient prediction method and generating a two-dimensional difference value histogram; selecting an optimal path in the two-dimensional difference histogram by adopting a dynamic programming algorithm, and establishing a first corresponding relation between a pixel pair corresponding to a difference value pair on the optimal path and bit information carried by a watermark to be embedded; if the difference value pair corresponding to the pixel pair is not on the optimal path, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is on the optimal path, modifying the pixel pair according to a preset first rule if the value of the bit information corresponding to the pixel pair is 1 according to the first corresponding relation; and if the value of the bit information corresponding to the pixel pair is 0, the pixel pair is not modified. The scheme has high data embedding capacity, low embedding distortion and high safety of hidden data.
Drawings
Fig. 1 is a schematic flowchart of an adaptive reversible watermarking method according to an embodiment of the present invention;
fig. 2 is a schematic neighborhood diagram of a pixel pair (x, y) in an adaptive reversible watermarking method according to an embodiment of the present invention;
fig. 3 is a flowchart illustrating an adaptive reversible watermarking method according to a second embodiment of the present invention;
fig. 4 is a schematic flowchart of an adaptive reversible watermarking method according to a third embodiment of the present invention;
fig. 5 is a schematic structural diagram of an adaptive reversible watermarking system according to a fourth embodiment of the present invention;
fig. 6 is a schematic structural diagram of an adaptive reversible watermarking system according to a fifth embodiment of the present invention;
fig. 7 is a schematic structural diagram of an adaptive reversible watermarking system according to a sixth embodiment of the present invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
Fig. 1 is a schematic flowchart of an adaptive reversible watermarking method according to an embodiment of the present invention, and as shown in fig. 1, the method includes the following steps:
11. dividing the carrier image into a plurality of pixel pairs (x, y) which are not overlapped with each other, wherein two pixel points in each pixel pair are adjacent, wherein x is the pixel value of the previous pixel point, and y is the pixel value of the next pixel point.
Specifically, the carrier image may be divided into a plurality of pixel pairs that do not overlap with each other in a plurality of ways, for example, two pixel points adjacent to each other in the horizontal direction of the carrier image are used as a group of pixel pairs, and the carrier image is divided into a plurality of pixel pairs that do not overlap with each other.
12. For each pixel pair (x, y), calculating a corresponding prediction value z according to the neighborhood of the pixel pair by adopting a gradient prediction method, and calculating a corresponding difference value pair (d) by utilizing a first formulax,dy) The first formula is:
d1＝x-z，d2＝y-z。
specifically, the neighborhood of the pixel pair (x, y) is as shown in fig. 2, and for each pixel pair (x, y), the calculation method for calculating the corresponding predicted value z by using the gradient prediction method according to the neighborhood of the pixel pair is as follows:
wherein v is1～v8Is the pixel value, d, of the corresponding pixel point in the neighborhood of the pixel pair (x, y)v、dhRepresenting the gradient of the neighborhood in the vertical direction and the horizontal direction, respectively, wherein:
dv＝|v1-v5|+|v3-v7|+|v4-v8|；
dh＝|v1-v2|+|v3-v4|+|v4-v5@；
13. and generating a two-dimensional difference value histogram according to the difference value pairs corresponding to the plurality of pixel pairs.
For example, in order to improve the security of embedding secret information, data embedding is not performed on the pixel points of the complex texture area of the carrier image, that is, the difference pairs corresponding to the pixel pairs of the complex texture area of the carrier image are not selected to generate the two-dimensional difference histogram, that is, the difference pairs corresponding to the pixel pairs of the smooth texture area of the image are selected to generate the two-dimensional difference histogram. On the basis of the embodiment shown in fig. 1, 13 may specifically include: and calculating the noise level of each pixel pair, and generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs with the noise levels smaller than the noise threshold value.
Further, for a pixel with a pixel value of 255 or 0 in the carrier image, the pixel cannot be modified by +1 or-1, which is pixel overflow. When selecting the difference pairs to generate the two-dimensional histogram, the difference pairs corresponding to the pixel pairs with pixel overflow may not be selected. As an optional implementation manner, on the basis of any of the above implementation manners, 13 may specifically include: counting the information of the difference value pairs corresponding to the pixel pairs with pixel overflow, and recording the information to a position mapping table L; and selecting the difference value pairs corresponding to the pixel pairs without pixel overflow to generate a two-dimensional difference value histogram.
The two aforementioned embodiments can also be combined to implement, for example, the hypothetical pixel pair (x) in the actual scenei,yi) The corresponding predicted value is z 124 (125 ), and the noise level is calculated to be NL (x)i,yi) 10, the corresponding difference value pair is (d)1,d2) If the noise threshold is T25, (1,1) the noise level of the difference value pair is less than the threshold, and there is no pixel overflow in the pixel pair, so it can be used to generate a two-dimensional difference histogram. Correspondingly, 13 may specifically include: calculating a noise level for each pixel pair; counting the information of the difference value pairs corresponding to the pixel pairs with pixel overflow, and recording the information to a position mapping table L; and selecting a difference value pair corresponding to a pixel pair of which the noise level is less than a threshold value and which does not have pixel overflow, and generating a two-dimensional difference value histogram.
14. Selecting a starting point and an end point in the two-dimensional difference histogram, and selecting an optimal path in a grid formed by the starting point and the end point by adopting a dynamic programming algorithm, wherein the optimal path is a path with the maximum sum of the frequencies of first difference values passing through all paths from the starting point to the end point in the grid, the actual embeddable capacity of a pixel pair corresponding to the first difference value pair passing through the optimal path is not less than the capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in the horizontal direction or the vertical direction.
Specifically, the dynamic planning algorithm is adopted, and the formal definition of the optimal path selected in the grid formed by the starting point and the end point is shown as the following formula:
D(x，y)＝max{D(x-1，y)，D(x，y-1)}+h(x，y)
wherein D (x, y) represents the sum of the frequencies of the difference pairs passed by the optimal embedding path with the end point of (x, y), and h (x, y) represents the number of the pixel pairs corresponding to the difference pair (x, y) in the carrier image, that is, the frequency of the difference pair (x, y). The formula can be solved in linear time complexity using a dynamic programming algorithm.
15. Defining the first difference pair as a dilated difference pair and the remaining difference pairs as translated difference pairs.
16. And establishing a first corresponding relation between the pixel pair corresponding to the expansion difference value pair and bit information carried by the watermark to be embedded.
17. For each pixel pair, if the difference value pair corresponding to the pixel pair is a translation difference value pair, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; and if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair.
As an optional implementation manner, on the basis of any one of the foregoing implementation manners, the modifying, according to the preset first rule, the pixel pair in 17 may specifically include:
moving the difference value pairs corresponding to the pixel pairs in the two-dimensional difference value histogram in a direction away from the optimal path according to the first rule, wherein the moved difference value pairs are not overlapped with each other, and the moving distance of the difference value pairs is a preset threshold value;
and calculating the pixel value of each pixel point in the current pixel pair according to the coordinate of the differential value pair corresponding to the moved pixel pair and the first formula, and modifying the pixel value of each pixel point in the pixel pair into the pixel value obtained by current calculation.
By the scheme, the watermark can be embedded in the carrier image. In particular, in order to reduce the embedding distortion of the carrier image, the preset threshold value may be defined as 1, i.e. the difference value pair (d)1,d2) May be (d)1-1,d2) Or (d) or1+1,d2) Or (d) or1,d2-1), or (d)1,d2+1), it can be calculated according to the first formula that the modification direction of the pixel pair corresponding to the difference value pair is (x-1, y), or (x +1, y), or (x, y-1), or (x, y +1), i.e. the modification amount of the pixel pair is 1.
Further, in order to be able to subsequently extract the bit information carried by the embedded watermark and recover the carrier image, the method may further include embedding additional information in the carrier image, the additional information including: the position information of the first difference value pair passed by the optimal path, the position information of the pixel pair corresponding to the last bit information of the bit information carried by the watermark to be embedded, the noise threshold, the position mapping table L and the length of the position mapping table L.
In the adaptive reversible watermarking method provided by this embodiment, a carrier image is divided into a plurality of non-overlapping adjacent pixel pairs; calculating corresponding difference value pairs by using a gradient prediction method and generating a two-dimensional difference value histogram; selecting an optimal path in the two-dimensional difference histogram by adopting a dynamic programming algorithm, and establishing a first corresponding relation between a pixel pair corresponding to a difference value pair on the optimal path and bit information carried by a watermark to be embedded; if the difference value pair corresponding to the pixel pair is not on the optimal path, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is on the optimal path, modifying the pixel pair according to a preset first rule if the value of the bit information corresponding to the pixel pair is 1 according to the first corresponding relation; and if the value of the bit information corresponding to the pixel pair is 0, the pixel pair is not modified. The scheme has high data embedding capacity, low embedding distortion and high safety of hidden data.
For example, a starting point and an end point are selected from the two-dimensional difference histogram in an actual application scene, and a dynamic programming algorithm is adopted, so that various methods can be used for selecting an optimal path in a grid formed by the starting point and the end point, for example, a grid with a preset side length can be drawn by taking the origin (0,0) of the two-dimensional difference histogram as a central point, and two difference value pairs on a diagonal line of the grid are taken as the starting point and the end point; alternatively, two difference value pairs can be arbitrarily selected as the starting point and the end point. Specifically, as shown in fig. 3, fig. 3 is a schematic flowchart of a method for adaptive reversible watermarking according to a second embodiment of the present invention, and on the basis of the implementation manner shown in fig. 1, 14 specifically includes:
341. selecting any two difference value pairs as a current starting point and a current end point;
342. selecting a current optimal path in a grid formed by the starting point and the end point according to a preset first rule by adopting the dynamic programming algorithm according to the current starting point and the current end point;
343. calculating the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path;
344. if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of bit information carried by the watermark to be embedded currently, re-selecting the starting point and the end point as the current starting point and the current end point; and returning to execute 342, wherein the step of selecting a current optimal path in a grid formed by the starting point and the end point according to a preset first rule by using the dynamic programming algorithm according to the current starting point and the end point until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not less than the capacity of the bit information carried by the current watermark to be embedded.
Specifically, in the above 344, there are various methods for re-selecting the start point and the end point as the current start point and end point, for example, the start point and the end point may be re-selected arbitrarily as the current start point and end point, or the center of the grid formed by the current start point and end point is used as the center point, the side length of the grid is enlarged by 2 times, and two difference value pairs on the diagonal line of the enlarged grid are selected as the current start point and end point.
Further, in the above 344, if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value that the current optimal path passes through is much larger than the capacity of the bit information carried by the watermark to be embedded currently, a bisection method may be adopted to reselect the start point and the end point as the current start point and end point, and the step of selecting the current optimal path in the grid formed by the start point and the end point according to the current start point and end point in the execution 342 is performed, by using the dynamic programming algorithm, according to a preset first rule, until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value that the current optimal path passes through is not smaller than and close to the capacity of the bit information carried by the watermark to be embedded currently.
In the self-adaptive reversible watermarking method provided by the embodiment, any two difference values are selected as the current starting point and the current end point, and a dynamic programming algorithm is adopted to select the current optimal path in a grid formed by the current starting point and the current end point according to a preset first rule; if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of bit information carried by the watermark to be embedded currently, re-selecting the starting point and the end point as the current starting point and the current end point; and returning to execute the step of selecting the current optimal path in the grid formed by the starting point and the end point by adopting a dynamic programming algorithm according to the current starting point and the current end point until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not less than the capacity of the bit information carried by the current watermark to be embedded. The scheme can quickly select the pixel pairs capable of data embedding, and the data embedding efficiency is high.
Fig. 4 is a schematic flow chart of an adaptive reversible watermarking method according to a third embodiment of the present invention, as shown in fig. 4, on the basis of the implementation manner shown in fig. 1, 14 specifically may include:
441. dividing the two-dimensional difference value histogram into four mutually disjoint subsets A according to a preset second rulenWherein n is 1,2,3, 4; the second rule is:
442. dividing the capacity of bit information carried by the watermark to be embedded into the capacity corresponding to each subset according to a preset third rule;
specifically, according to a preset third rule, there may be multiple methods for dividing the capacity of the bit information carried by the watermark to be embedded into the capacity corresponding to each subset, for example, the capacity of the bit information carried by the watermark to be embedded may be equally divided into the capacity corresponding to each subset.
443. For each subset, selecting a difference value pair closest to (0,0) in the subset as a starting point corresponding to the subset, and selecting an end point corresponding to the subset; selecting an optimal path corresponding to the subset in a grid formed by a starting point corresponding to the subset and an end point corresponding to the subset by adopting a dynamic programming algorithm, wherein the optimal path corresponding to the subset is a path with the maximum sum of frequencies of passing difference values in all paths from the starting point corresponding to the subset to the end point corresponding to the subset in the grid, the actual embeddable capacity of a pixel pair corresponding to the passing difference value pair of the optimal path corresponding to the subset is not less than the capacity corresponding to the subset, and adjacent difference values passing through the optimal path corresponding to the subset are adjacent in the horizontal direction or the vertical direction;
444. and adding the optimal paths corresponding to each subset to obtain the optimal paths.
As an optional implementation manner, on the basis of the implementation manner shown in fig. 4, the selecting an endpoint corresponding to the subset in 443 may specifically include:
4431. selecting any one difference value pair in the subset as a current terminal;
4432. selecting a current optimal path of the subset in a grid formed by the starting point and the end point according to a preset first rule by adopting the dynamic programming algorithm according to the starting point and the current end point of the subset;
4433. calculating the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the difference value pair passed by the current optimal path of the subset;
4434. if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the difference value pair passed by the current optimal path of the subset is smaller than the capacity corresponding to the subset, reselecting the terminal point from the subset as the current terminal point; and returning to execute the step of selecting the current optimal path of the subset in the grid formed by the starting point and the end point by adopting the dynamic planning algorithm according to the starting point and the current end point of the subset until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the differential value passed by the current optimal path of the subset is not less than the capacity corresponding to the subset.
Specifically, in 4434, there may be a plurality of methods for re-selecting the end point in the subset as the current end point, for example, the end point may be randomly selected in the subset as the current end point, or the side length of the grid formed by the start point of the subset and the current end point is enlarged by 2 times, and another difference value pair on a diagonal line where the start point of the subset is located on the enlarged grid is selected as the current end point.
Further, in 4434, if the actual capacity of the pixel pair corresponding to the differential value pair passed by the current optimal path of the subset is much larger than the capacity corresponding to the subset, a bisection method may be adopted to reselect the end point as the current end point, and the step of selecting the current optimal path of the subset in the grid formed by the start point and the end point according to the preset first rule by using the dynamic programming algorithm and returning to 4432 is executed until the actual capacity of the pixel pair corresponding to the differential value pair passed by the current optimal path of the subset is not smaller than and close to the capacity corresponding to the subset.
In the adaptive reversible watermarking method provided by this embodiment, the two-dimensional difference histogram is divided into four mutually disjoint subsets according to a preset second rule, and the capacity of bit information carried by a watermark to be embedded is divided into a capacity corresponding to each subset according to a preset third rule; and aiming at each subset, selecting an optimal path corresponding to each subset by adopting a dynamic programming algorithm, and adding the optimal paths corresponding to each subset to obtain the optimal path of the whole two-dimensional difference histogram. The sum of the frequencies of the first difference value pairs passed by the optimal path obtained by the scheme is maximum, so that higher data embedding capacity and smaller embedding distortion can be obtained.
Fig. 5 is a schematic structural diagram of an adaptive reversible watermarking system according to a fourth embodiment of the present invention, and as shown in fig. 5, the system includes:
the pixel pair module 51 is configured to divide the carrier image into a plurality of pixel pairs (x, y) that do not coincide with each other, where two pixel points in each pixel pair are adjacent to each other, where x is a pixel value of a previous pixel point, and y is a pixel value of a subsequent pixel point.
Specifically, the carrier image may be divided into a plurality of pixel pairs that do not overlap with each other in a plurality of ways, for example, two pixel points adjacent to each other in the horizontal direction of the carrier image are used as a group of pixel pairs, and the carrier image is divided into a plurality of pixel pairs that do not overlap with each other.
A difference value pair module 52, configured to, for each pixel pair (x, y), calculate a corresponding prediction value z by using a gradient prediction method according to a neighborhood of the pixel pair, and calculate a corresponding difference value pair (d) by using a first formulax,dy) The first formula is:
d1＝x-z，d2＝y-z。
the histogram module 53 is configured to generate a two-dimensional difference histogram according to the difference pairs corresponding to the plurality of pixel pairs.
For example, in order to improve the security of embedding secret information, data embedding is not performed on the pixel points of the complex texture area of the carrier image, that is, the difference pairs corresponding to the pixel pairs of the complex texture area of the carrier image are not selected to generate the two-dimensional difference histogram, that is, the difference pairs corresponding to the pixel pairs of the smooth texture area of the image are selected to generate the two-dimensional difference histogram. On the basis of the embodiment shown in fig. 5, the histogram module 53 may include: and the noise unit is used for calculating the noise level of each pixel pair and generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs with the noise levels smaller than the noise threshold value.
Further, for a pixel with a pixel value of 255 or 0 in the carrier image, the pixel cannot be modified by +1 or-1, which is pixel overflow. When selecting the difference pairs to generate the two-dimensional histogram, the difference pairs corresponding to the pixel pairs with pixel overflow may not be selected. As an optional implementation manner, on the basis of any of the above implementation manners, the histogram module 53 may include: the overflow unit is used for counting the information of the difference value pairs corresponding to the pixel pairs with pixel overflow and recording the information to the position mapping table L; and selecting the difference value pairs corresponding to the pixel pairs without pixel overflow to generate a two-dimensional difference value histogram.
Correspondingly, on the basis of the processing of the noise unit and the overflow unit, the histogram module 53 is specifically configured to select a difference pair corresponding to a pixel pair whose noise level is less than a threshold and whose pixel overflow does not exist, and generate a two-dimensional difference histogram.
An optimal path module 54, configured to select a starting point and an end point in the two-dimensional difference histogram, and select an optimal path in a grid formed by the starting point and the end point by using a dynamic programming algorithm, where the optimal path is a path in which a sum of frequencies of first difference values passing through all paths from the starting point to the end point in the grid is the largest, an actual embeddable capacity of a pixel pair corresponding to the first difference value passing through the optimal path is not less than a capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in a horizontal direction or a vertical direction.
Specifically, the dynamic planning algorithm is adopted, and the formal definition of the optimal path selected in the grid formed by the starting point and the end point is shown as the following formula:
D(x，y)＝max{D(x-1，y)，D(x，y-1)}+h(x，y)
wherein D (x, y) represents the sum of the frequencies of the difference pairs passed by the optimal embedding path with the end point of (x, y), and h (x, y) represents the number of the pixel pairs corresponding to the difference pair (x, y) in the carrier image, that is, the frequency of the difference pair (x, y). The formula can be solved in linear time complexity using a dynamic programming algorithm.
A defining module 55 is configured to define the first difference pair as an expansion difference pair, and define the remaining difference pairs as translation difference pairs.
A corresponding module 56, configured to establish a first corresponding relationship between the pixel pair corresponding to the expanded difference value pair and bit information carried by the watermark to be embedded.
The modifying module 57 is configured to, for each pixel pair, modify the pixel pair according to a preset first rule if the difference pair corresponding to the pixel pair is a translational difference pair; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; and if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair.
As an alternative implementation, on the basis of any of the above embodiments, the modification module 57 may include:
a moving unit, configured to move, according to the first rule, the difference pairs corresponding to the pixel pairs in the two-dimensional difference histogram in a direction away from the optimal path, where the moved difference pairs are not overlapped with each other, and a moving distance of the difference pair is a preset threshold;
and the modifying unit is used for calculating the pixel value of each pixel point in the current pixel pair according to the coordinate of the difference value pair corresponding to the moved pixel pair and the first formula, and modifying the pixel value of each pixel point in the current pixel pair into the currently calculated pixel value.
By the scheme, the watermark can be embedded in the carrier image. In particular, in order to reduce the embedding distortion of the carrier image, the preset threshold value may be defined as 1, i.e. the difference value pair (d)1,d2) May be (d)1-1,d2) Or (d) or1+1,d2) Or (d) or1,d2-1), or (d)1，d2+1), it can be calculated according to the first formula that the modification direction of the pixel pair corresponding to the difference value pair is (x-1, y), or (x +1, y), or (x, y-1), or (x, y +1), i.e. the modification amount of the pixel pair is 1.
Further, in order to be able to extract the bit information carried by the embedded watermark and recover the carrier image subsequently, the system may further include an additional module for embedding additional information in the carrier image, where the additional information includes: the position information of the first difference value pair passed by the optimal path, the position information of the pixel pair corresponding to the last bit information of the bit information carried by the watermark to be embedded, the noise threshold, the position mapping table L and the length of the position mapping table L.
In the adaptive reversible watermark system provided by this embodiment, the carrier image is divided into a plurality of non-overlapping adjacent pixel pairs; calculating corresponding difference value pairs by using a gradient prediction method and generating a two-dimensional difference value histogram; selecting an optimal path in the two-dimensional difference histogram by adopting a dynamic programming algorithm, and establishing a first corresponding relation between a pixel pair corresponding to a difference value pair on the optimal path and bit information carried by a watermark to be embedded; if the difference value pair corresponding to the pixel pair is not on the optimal path, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is on the optimal path, modifying the pixel pair according to a preset first rule if the value of the bit information corresponding to the pixel pair is 1 according to the first corresponding relation; and if the value of the bit information corresponding to the pixel pair is 0, the pixel pair is not modified. The scheme has high data embedding capacity, low embedding distortion and high safety of hidden data.
For example, a starting point and an end point are selected from the two-dimensional difference histogram in an actual application scene, and a dynamic programming algorithm is adopted, so that various methods can be used for selecting an optimal path in a grid formed by the starting point and the end point, for example, a grid with a preset side length can be drawn by taking the origin (0,0) of the two-dimensional difference histogram as a central point, and two difference value pairs on a diagonal line of the grid are taken as the starting point and the end point; alternatively, two difference value pairs can be arbitrarily selected as the starting point and the end point. Specifically, as shown in fig. 6, fig. 6 is a schematic structural diagram of an adaptive reversible watermarking system according to a fifth embodiment of the present invention, and based on the implementation shown in fig. 5, the optimal path module 54 may include:
the first selecting unit 641 is configured to select any two difference pairs as a current starting point and a current end point;
the optimal path unit 642 is configured to select, according to the current starting point and the current end point, the current optimal path in the grid formed by the starting point and the end point according to a preset first rule by using the dynamic planning algorithm;
a calculating unit 643, configured to calculate an actual capacity of an actual embeddable capacity of the pixel pair corresponding to the first difference value pair that the current optimal path passes through;
a second selecting unit 644, configured to reselect the starting point and the end point as the current starting point and end point if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of the bit information carried by the watermark to be embedded currently; and returning to execute 342, wherein the step of selecting a current optimal path in a grid formed by the starting point and the end point according to a preset first rule by using the dynamic programming algorithm according to the current starting point and the end point until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not less than the capacity of the bit information carried by the current watermark to be embedded.
Specifically, the second selecting unit 644 may reselect the start point and the end point as the current start point and end point, for example, the start point and the end point may be reselected as the current start point and end point, or the center of the grid formed by the current start point and end point is used as the center point, the side length of the grid is enlarged by 2 times, and two difference values on the diagonal line of the enlarged grid are selected as the current start point and end point.
Further, the second selecting unit 644 may be further configured to, if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is far greater than the capacity of the bit information carried by the current watermark to be embedded, reselect the start point and the end point by using a bisection method as the current start point and the end point, and instruct the optimal path unit 642 to execute the step of selecting the current optimal path in the grid formed by the start point and the end point again according to the current start point and the current end point and by using the dynamic programming algorithm and according to a preset first rule until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not smaller than and approaches the capacity of the bit information carried by the current watermark to be embedded.
In the adaptive reversible watermark system provided by this embodiment, any two difference values are selected as the current starting point and end point, and a dynamic programming algorithm is adopted to select the current optimal path in the grid formed by the current starting point and end point according to a preset first rule; if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of bit information carried by the watermark to be embedded currently, re-selecting the starting point and the end point as the current starting point and the current end point; and returning to execute the step of selecting the current optimal path in the grid formed by the starting point and the end point by adopting a dynamic programming algorithm according to the current starting point and the current end point until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not less than the capacity of the bit information carried by the current watermark to be embedded. The scheme can quickly select the pixel pairs capable of data embedding, and the data embedding efficiency is high.
Fig. 7 is a schematic structural diagram of an adaptive reversible watermarking system according to a sixth embodiment of the present invention, and as shown in fig. 7, on the basis of the implementation shown in fig. 5, the optimal path module 54 may include:
a first dividing unit 741, configured to divide the two-dimensional difference histogram into four mutually disjoint subsets a according to a preset second rulenWherein n is 1,2,3, 4; the second ruleComprises the following steps:
a second dividing unit 742 is configured to divide the capacity of the bit information carried by the watermark to be embedded into the capacity corresponding to each subset according to a preset third rule;
specifically, according to a preset third rule, there may be multiple methods for dividing the capacity of the bit information carried by the watermark to be embedded into the capacity corresponding to each subset, for example, the capacity of the bit information carried by the watermark to be embedded may be equally divided into the capacity corresponding to each subset.
A subset unit 743, configured to, for each of the subsets, select a difference pair closest to (0,0) in the subset as a starting point corresponding to the subset, and select an end point corresponding to the subset; selecting an optimal path corresponding to the subset in a grid formed by a starting point corresponding to the subset and an end point corresponding to the subset by adopting a dynamic programming algorithm, wherein the optimal path corresponding to the subset is a path with the maximum sum of frequencies of passing difference values in all paths from the starting point corresponding to the subset to the end point corresponding to the subset in the grid, the actual embeddable capacity of a pixel pair corresponding to the passing difference value pair of the optimal path corresponding to the subset is not less than the capacity corresponding to the subset, and adjacent difference values passing through the optimal path corresponding to the subset are adjacent in the horizontal direction or the vertical direction;
and the combining unit 744 is configured to add the optimal paths corresponding to each subset to obtain the optimal path.
As an alternative implementation, on the basis of the implementation shown in fig. 7, the subset unit 743 may include:
the first selection subunit is used for selecting any one difference value pair in the subset as a current terminal;
the optimal path subunit is used for selecting a current optimal path of the subset in a grid formed by the starting point and the end point according to a preset first rule by adopting the dynamic planning algorithm according to the starting point and the current end point of the subset;
the calculating subunit is used for calculating the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the difference value pair passed by the current optimal path of the subset;
the second selecting subunit is used for re-selecting the end point in the subset as the current end point if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the difference value passed by the current optimal path of the subset is smaller than the capacity corresponding to the subset; and returning to execute the step of selecting the current optimal path of the subset in the grid formed by the starting point and the end point by adopting the dynamic planning algorithm according to the starting point and the current end point of the subset until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the differential value passed by the current optimal path of the subset is not less than the capacity corresponding to the subset.
Specifically, the second selecting subunit may reselect the end point in the subset as the current end point in a variety of ways, for example, the end point may be randomly reselected in the subset as the current end point, or the side length of the grid formed by the start point of the subset and the current end point is enlarged by 2 times, and another difference value pair on a diagonal where the start point of the subset is located on the enlarged grid is selected as the current end point.
Further, the second selecting subunit may be further configured to, if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the differential value pair passed by the current optimal path of the subset is much larger than the capacity corresponding to the subset, adopt a bisection method to reselect the end point as the current end point, and instruct the optimal path subunit to execute the step of selecting the current optimal path of the subset in the grid formed by the start point and the end point again according to the start point and the current end point of the subset by using the dynamic programming algorithm and according to a preset first rule until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the differential value pair passed by the current optimal path of the subset is not smaller than and close to the capacity corresponding to the subset.
In the adaptive reversible watermark system provided in this embodiment, the two-dimensional difference histogram is divided into four mutually disjoint subsets according to a preset second rule, and the capacity of bit information carried by a watermark to be embedded is divided into a capacity corresponding to each subset according to a preset third rule; and aiming at each subset, selecting an optimal path corresponding to each subset by adopting a dynamic programming algorithm, and adding the optimal paths corresponding to each subset to obtain the optimal path of the whole two-dimensional difference histogram. The sum of the frequencies of the first difference value pairs passed by the optimal path obtained by the scheme is maximum, so that higher data embedding capacity and smaller embedding distortion can be obtained.
It can be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working process of the system described above may refer to the corresponding process in the foregoing method embodiment, and is not described herein again.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (8)
1. An adaptive reversible watermarking method, comprising:
dividing the carrier image into a plurality of pixel pairs (x, y) which are not overlapped with each other, wherein two pixel points in each pixel pair are adjacent, wherein x is the pixel value of the previous pixel point, and y is the pixel value of the next pixel point;
for each pixel pair (x, y), calculating a corresponding prediction value z according to the neighborhood of the pixel pair by adopting a gradient prediction method, and calculating a corresponding difference value pair (d) by utilizing a first formulax,dy) The first formula is:
d1＝x-z，d2＝y-z；
generating a two-dimensional difference value histogram according to the difference value pairs corresponding to the plurality of pixel pairs;
selecting a starting point and an end point in the two-dimensional difference histogram, and selecting an optimal path in a grid formed by the starting point and the end point by adopting a dynamic programming algorithm, wherein the optimal path is a path with the maximum sum of the frequencies of first difference values passing through all paths from the starting point to the end point in the grid, the actual embeddable capacity of a pixel pair corresponding to the first difference value pair passing through the optimal path is not less than the capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in the horizontal direction or the vertical direction;
defining the first difference value pair as an expansion difference value pair, and defining the rest difference value pairs as translation difference value pairs;
establishing a first corresponding relation between the pixel pair corresponding to the expansion difference value pair and bit information carried by the watermark to be embedded;
for each pixel pair, if the difference value pair corresponding to the pixel pair is a translation difference value pair, modifying the pixel pair according to a preset first rule; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair;
the modifying the pixel pair according to the preset first rule comprises: moving the difference value pairs corresponding to the pixel pairs in the two-dimensional difference value histogram in a direction away from the optimal path according to the first rule, wherein the moved difference value pairs are not overlapped with each other, and the moving distance of the difference value pairs is a preset threshold value; and calculating the pixel value of each pixel point in the current pixel pair according to the coordinate of the differential value pair corresponding to the moved pixel pair and the first formula, and modifying the pixel value of each pixel point in the pixel pair into the pixel value obtained by current calculation.
2. The method of claim 1, wherein generating a two-dimensional difference histogram from the difference pairs corresponding to the plurality of pixel pairs comprises:
and calculating the noise level of each pixel pair, and generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs with the noise levels smaller than the noise threshold value.
3. The method according to claim 1, wherein the selecting a starting point and an end point in the two-dimensional difference histogram, and selecting an optimal path within a grid formed by the starting point and the end point by using a dynamic programming algorithm, comprises:
selecting any two difference value pairs as a current starting point and a current end point;
selecting a current optimal path in a grid formed by the starting point and the end point according to a preset first rule by adopting the dynamic programming algorithm according to the current starting point and the current end point;
calculating the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path;
if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of bit information carried by the watermark to be embedded currently, re-selecting the starting point and the end point as the current starting point and the current end point; and returning to execute the step of selecting the current optimal path in the grid formed by the starting point and the end point according to the current starting point and the current end point by adopting the dynamic programming algorithm and according to a preset first rule until the actual capacity of the actual embeddable capacity of the corresponding pixel pair is not less than the capacity of the bit information carried by the current watermark to be embedded.
4. The method according to claim 1, wherein the selecting a starting point and an end point in the two-dimensional difference histogram, and selecting an optimal path within a grid formed by the starting point and the end point by using a dynamic programming algorithm, comprises:
dividing the two-dimensional difference value histogram into four mutually disjoint subsets A according to a preset second rulenWherein n is 1,2,3, 4; the second rule is:
dividing the capacity of bit information carried by the watermark to be embedded into the capacity corresponding to each subset according to a preset third rule;
for each subset, selecting a difference value pair closest to (0,0) in the subset as a starting point corresponding to the subset, and selecting an end point corresponding to the subset; selecting an optimal path corresponding to the subset in a grid formed by a starting point corresponding to the subset and an end point corresponding to the subset by adopting a dynamic programming algorithm, wherein the optimal path corresponding to the subset is a path with the maximum sum of frequencies of passing difference values in all paths from the starting point corresponding to the subset to the end point corresponding to the subset in the grid, the actual embeddable capacity of a pixel pair corresponding to the passing difference value pair of the optimal path corresponding to the subset is not less than the capacity corresponding to the subset, and adjacent difference values passing through the optimal path corresponding to the subset are adjacent in the horizontal direction or the vertical direction;
and adding the optimal paths corresponding to each subset to obtain the optimal paths.
5. An adaptive reversible watermarking system, comprising:
the image processing device comprises a pixel pair module, a pixel value acquisition module and a pixel value acquisition module, wherein the pixel pair module is used for dividing a carrier image into a plurality of pixel pairs (x, y) which are not overlapped with each other, and two pixel points in each pixel pair are adjacent, wherein x is the pixel value of a previous pixel point, and y is the pixel value of a next pixel point;
difference valueA pair module, configured to, for each pixel pair (x, y), calculate a corresponding prediction value z according to a neighborhood of the pixel pair by using a gradient prediction method, and calculate a corresponding difference value pair (d) by using a first formulax,dy) The first formula is:
d1＝x-z，d2＝y-z；
the histogram module is used for generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs;
an optimal path module, configured to select a starting point and an end point in the two-dimensional difference histogram, and select an optimal path in a grid formed by the starting point and the end point by using a dynamic programming algorithm, where the optimal path is a path in which a sum of frequencies of first difference values passing through all paths from the starting point to the end point in the grid is the largest, an actual embeddable capacity of a pixel pair corresponding to the first difference value passing through the optimal path is not less than a capacity of bit information carried by a watermark to be embedded, and adjacent difference values passing through the optimal path are adjacent in a horizontal direction or a vertical direction;
a defining module, configured to define the first difference pair as an expansion difference pair, and define the remaining difference pairs as translation difference pairs;
a corresponding module, configured to establish a first corresponding relationship between the pixel pair corresponding to the expanded difference value pair and bit information carried by the watermark to be embedded;
the modification module is used for modifying the pixel pair according to a preset first rule if the difference value pair corresponding to the pixel pair is a translation difference value pair aiming at each pixel pair; if the difference value pair corresponding to the pixel pair is an expansion difference value pair, modifying the pixel pair according to a preset first rule according to the first corresponding relation if the value of the bit information corresponding to the pixel pair is 1; if the value of the bit information corresponding to the pixel pair is 0, not modifying the pixel pair;
the modification module includes: a mobile unit and a modification unit;
the moving unit is configured to move the difference pairs corresponding to the pixel pairs in the two-dimensional difference histogram in a direction away from the optimal path according to the first rule, where the moved difference pairs are not overlapped with each other, and a moving distance of the difference pair is a preset threshold;
and the modification unit is used for calculating the pixel value of each pixel point in the current pixel pair according to the coordinate of the difference value pair corresponding to the moved pixel pair and the first formula, and modifying the pixel value of each pixel point in the current pixel pair into the currently calculated pixel value.
6. The system of claim 5, wherein the histogram module comprises:
and the noise unit is used for calculating the noise level of each pixel pair and generating a two-dimensional difference histogram according to the difference value pairs corresponding to the plurality of pixel pairs with the noise levels smaller than the noise threshold value.
7. The system of claim 5, wherein the optimal path module comprises:
the first selecting unit is used for selecting any two difference value pairs as a current starting point and a current end point;
the optimal path unit is used for selecting a current optimal path in a grid formed by the starting point and the end point by adopting the dynamic planning algorithm according to the current starting point and the current end point;
the calculating unit is used for calculating the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path;
the second selection unit is used for reselecting the starting point and the end point as the current starting point and the current end point if the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value pair passed by the current optimal path is smaller than the capacity of bit information carried by the watermark to be embedded currently; and instructing the optimal path unit to execute the step of selecting the current optimal path in the grid formed by the starting point and the end point according to the current starting point and the current end point by adopting the dynamic programming algorithm and according to a preset first rule until the actual capacity of the actual embeddable capacity of the pixel pair corresponding to the first difference value passed by the current optimal path is not less than the capacity of the bit information carried by the current watermark to be embedded.
8. The system of claim 5, wherein the optimal path module comprises:
a first dividing unit, configured to divide the two-dimensional difference histogram into four mutually disjoint subsets a according to a preset second rulenWherein n is 1,2,3, 4; the second rule is:
a second dividing unit, configured to divide a capacity of bit information carried by the watermark to be embedded into a capacity corresponding to each subset according to a preset third rule;
a subset unit, configured to select, for each of the subsets, a difference pair closest to (0,0) in the subset as a starting point corresponding to the subset, and select an end point corresponding to the subset; selecting an optimal path corresponding to the subset in a grid formed by a starting point corresponding to the subset and an end point corresponding to the subset by adopting a dynamic programming algorithm, wherein the optimal path corresponding to the subset is a path with the maximum sum of frequencies of passing difference values in all paths from the starting point corresponding to the subset to the end point corresponding to the subset in the grid, the actual embeddable capacity of a pixel pair corresponding to the passing difference value pair of the optimal path corresponding to the subset is not less than the capacity corresponding to the subset, and adjacent difference values passing through the optimal path corresponding to the subset are adjacent in the horizontal direction or the vertical direction;
and the combination unit is used for adding the optimal paths corresponding to each subset to obtain the optimal paths.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710326906.6A CN108876691B (en) | 2017-05-10 | 2017-05-10 | Self-adaptive reversible watermarking method and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710326906.6A CN108876691B (en) | 2017-05-10 | 2017-05-10 | Self-adaptive reversible watermarking method and system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108876691A CN108876691A (en) | 2018-11-23 |
| CN108876691B true CN108876691B (en) | 2022-03-08 |
Family
ID=64319108
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710326906.6A Expired - Fee Related CN108876691B (en) | 2017-05-10 | 2017-05-10 | Self-adaptive reversible watermarking method and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108876691B (en) |
Families Citing this family (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110232651B (en) * | 2019-06-12 | 2021-06-01 | 山东科技大学 | Reversible information hiding method based on block prediction and differential value histogram translation |
| CN115695673B (en) * | 2022-12-30 | 2023-04-21 | 中国科学技术大学 | Image anti-steganography method and device based on joint distortion |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0926882A3 (en) * | 1997-12-26 | 2002-08-21 | Sharp Kabushiki Kaisha | An image forming method and an image forming apparatus therefor |
| US7006662B2 (en) * | 2001-12-13 | 2006-02-28 | Digimarc Corporation | Reversible watermarking using expansion, rate control and iterative embedding |
| US7508552B2 (en) * | 2003-06-06 | 2009-03-24 | Konica Minolta Business Technologies, Inc. | Document reader, image forming apparatus, and image processing method |
| CN102567942A (en) * | 2012-01-09 | 2012-07-11 | 中山大学 | Reversible watermark method utilizing histogram shifting |
| CN102609896B (en) * | 2012-02-17 | 2014-03-26 | 中山大学 | Reversible watermark embedding and extracting method based on middle value keeping of histogram |
| CN102760280B (en) * | 2012-06-15 | 2014-12-17 | 苏州工业职业技术学院 | High-capacity reversible watermark embedding and extracting method as well as implement system thereof |
| CN106127670A (en) * | 2016-07-26 | 2016-11-16 | 广东工业大学 | A kind of processing method of reversible water mark image |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP3958128B2 (en) * | 2002-06-28 | 2007-08-15 | キヤノン株式会社 | Image processing apparatus and method, computer program, and computer-readable storage medium |
| US8818020B2 (en) * | 2009-09-18 | 2014-08-26 | Nanyang Technological University | Method of providing security for transmitting a digital medical image |
-
2017
- 2017-05-10 CN CN201710326906.6A patent/CN108876691B/en not_active Expired - Fee Related
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP0926882A3 (en) * | 1997-12-26 | 2002-08-21 | Sharp Kabushiki Kaisha | An image forming method and an image forming apparatus therefor |
| US7006662B2 (en) * | 2001-12-13 | 2006-02-28 | Digimarc Corporation | Reversible watermarking using expansion, rate control and iterative embedding |
| US7508552B2 (en) * | 2003-06-06 | 2009-03-24 | Konica Minolta Business Technologies, Inc. | Document reader, image forming apparatus, and image processing method |
| CN102567942A (en) * | 2012-01-09 | 2012-07-11 | 中山大学 | Reversible watermark method utilizing histogram shifting |
| CN102609896B (en) * | 2012-02-17 | 2014-03-26 | 中山大学 | Reversible watermark embedding and extracting method based on middle value keeping of histogram |
| CN102760280B (en) * | 2012-06-15 | 2014-12-17 | 苏州工业职业技术学院 | High-capacity reversible watermark embedding and extracting method as well as implement system thereof |
| CN106127670A (en) * | 2016-07-26 | 2016-11-16 | 广东工业大学 | A kind of processing method of reversible water mark image |
Non-Patent Citations (1)
| Title |
|---|
| "Improved reversible data hiding based on two-dimensional difference-histogram modification";Bowen Xue etc;《Springer》;20160716;13473-13491页 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108876691A (en) | 2018-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Kim et al. | Skewed histogram shifting for reversible data hiding using a pair of extreme predictions | |
| Qin et al. | An inpainting-assisted reversible steganographic scheme using a histogram shifting mechanism | |
| Arham et al. | Multiple layer data hiding scheme based on difference expansion of quad | |
| Qi et al. | Optimal reversible data hiding scheme based on multiple histograms modification | |
| Lu et al. | Secure halftone image steganography based on pixel density transition | |
| US9721318B2 (en) | Method for watermarking a three dimensional object and method for obtaining a payload from a three dimensional object | |
| CN107578362B (en) | Method for embedding secret information in image based on texture synthesis | |
| JP6543764B2 (en) | Security authentication method for digital video content and system therefor | |
| Lee et al. | Reversible data hiding using Lagrange interpolation for prediction-error expansion embedding | |
| Al Huti et al. | Increasing the capacity of the secret data using DEpixels blocks and adjusted RDE-based on grayscale images | |
| CN115131188A (en) | Robust image watermarking method based on generation countermeasure network | |
| CN108876691B (en) | Self-adaptive reversible watermarking method and system | |
| Lee et al. | Overlapping pixel value ordering predictor for high-capacity reversible data hiding | |
| Wang et al. | An improved reversible watermarking scheme using weighted prediction and watermarking simulation | |
| Zhang et al. | Real-time reversible data hiding based on multiple histogram modification | |
| Yüzkollar et al. | Region based interpolation error expansion algorithm for reversible image watermarking | |
| Wang | An efficient multiple-bit reversible data hiding scheme without shifting | |
| Luo et al. | Reversible image water marking based on prediction-error expansion and compensation | |
| Li et al. | Smaller Is Bigger: Rethinking the Embedding Rate of Deep Hiding | |
| CN111127288B (en) | Reversible image watermarking method, reversible image watermarking device and computer readable storage medium | |
| Isnanto et al. | Robustness of steganography image method using dynamic management position of least significant bit (LSB) | |
| Qu et al. | Reversible watermarking based on compensation | |
| CN112966230A (en) | Information steganography and extraction method, device and equipment | |
| Amsaveni et al. | An efficient reversible data hiding approach for colour images based on Gaussian weighted prediction error expansion and genetic algorithm | |
| Yang et al. | High-fidelity lossless data hiding based on predictors selection |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20220308 |