CN109118417A - Omnidirection context-prediction method and device towards reversible watermarking algorithm - Google Patents
Omnidirection context-prediction method and device towards reversible watermarking algorithm Download PDFInfo
- Publication number
- CN109118417A CN109118417A CN201810781894.0A CN201810781894A CN109118417A CN 109118417 A CN109118417 A CN 109118417A CN 201810781894 A CN201810781894 A CN 201810781894A CN 109118417 A CN109118417 A CN 109118417A
- Authority
- CN
- China
- Prior art keywords
- pixel
- subgraphs
- context
- subgraph
- prediction
- 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.)
- Withdrawn
Links
- 238000000034 method Methods 0.000 title claims abstract description 24
- 230000002441 reversible effect Effects 0.000 title claims abstract description 23
- 238000003780 insertion Methods 0.000 claims abstract description 55
- 230000037431 insertion Effects 0.000 claims abstract description 55
- 238000012545 processing Methods 0.000 claims abstract description 47
- 230000006835 compression Effects 0.000 description 5
- 238000007906 compression Methods 0.000 description 5
- 238000010586 diagram Methods 0.000 description 2
- 238000005516 engineering process Methods 0.000 description 2
- 238000012805 post-processing Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000007812 deficiency Effects 0.000 description 1
- 235000013399 edible fruits Nutrition 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000000465 moulding Methods 0.000 description 1
- 238000007639 printing Methods 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration using two or more images, e.g. averaging or subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Image Processing (AREA)
Abstract
The omnidirection context-prediction method and device towards reversible watermarking algorithm that the invention discloses a kind of, by constructing the fallout predictor based on omnidirection context, then original image is decomposed into four subgraphs, the prediction error of pixel in four subgraphs is calculated separately using fallout predictor, then watermark insertion processing is successively carried out to four subgraphs according to prediction error, it will be finally combined by four subgraphs of watermark insertion processing, form the final image for completing watermark insertion processing.Since other pixel predicted composition contexts all around current pixel in omnidirection context-prediction method of the invention, are utilized, and the value of current pixel is predicted by the prediction context, therefore can be greatly improved the precision of prediction.
Description
Technical field
The present invention relates to reversible Image Watermarking Technique field, especially a kind of omnidirection towards reversible watermarking algorithm
Context-prediction method and device.
Background technique
Reversible water mark is that the research hotspot of current digital watermark technology can be against the current compared with traditional digital watermark technology
Print can restore original hosted information completely without distortions, have biggish researching value and good application prospect, especially right
The fidelity of original hosted information requires high application field, such as the application fields such as information acquisition of taking photo by plane.To image into
When the reversible image watermark insertion of row is handled, need to carry out image compression processing, and when carrying out compression processing to image, often
It will use fallout predictor, but since in compression of images, the pixel of previous processed knows nothing the pixel value of post-processing, therefore,
Existing fallout predictor can only utilize the processed pixel in front, that is, current pixel top and left pixel, come
Predicted composition context, which limits the precision of predictions of fallout predictor.
Summary of the invention
For overcome the deficiencies in the prior art, the purpose of the present invention is to provide a kind of towards reversible watermarking algorithm
Omnidirection context-prediction method and device can utilize other pixel predicted composition contexts around current pixel, and right
The value of current pixel is predicted, to greatly improve the precision of prediction.
Technical solution used by the present invention solves the problems, such as it is:
Omnidirection context-prediction method towards reversible watermarking algorithm, comprising the following steps:
S1, fallout predictor of the building based on omnidirection context；
S2, original image is decomposed into four subgraphs, calculates separately the pre- of pixel in four subgraphs using fallout predictor
Survey error；
S3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
S4, it will be combined by four subgraphs of watermark insertion processing, formed and complete the final of watermark insertion processing
Image.
Further, the fallout predictor based on omnidirection context, the formula of fallout predictor are constructed in step S1 are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in
The pixel of pixel x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor immediately below pixel x [i, j]
Pixel.
Further, original image is decomposed into four subgraphs in step S2, original image be I=x [i, j] | 1≤i≤
H, 1≤j≤W }, wherein H and W is respectively the height and width of original image, and four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H'
Meet the following conditions respectively with W':
Further, the prediction error of pixel in four subgraphs is calculated separately in step S2 using fallout predictor, predicts error
It is acquired by following formula:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is son
The prediction error of pixel in image.
Further, watermark insertion processing is successively carried out to four subgraphs according to prediction error in step S3, including following
Step:
S31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to original image
Non-border pixel；
S32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
A kind of device storing the omnidirection context-prediction method towards reversible watermarking algorithm, including control module
With the storage medium for storing control instruction, the control module read the control instruction in the storage medium and execute with
Lower step:
Q1, fallout predictor of the building based on omnidirection context；
Q2, original image is decomposed into four subgraphs, calculates separately the pre- of pixel in four subgraphs using fallout predictor
Survey error；
Q3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
Q4, it will be combined by four subgraphs of watermark insertion processing, formed and complete the final of watermark insertion processing
Image.
Further, when control module executes step Q1, the fallout predictor based on omnidirection context, the formula of fallout predictor are constructed
Are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in
The pixel of pixel x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor immediately below pixel x [i, j]
Pixel.
Further, when control module executes step Q2, original image is decomposed into four subgraphs, original image I=
X [i, j] | 1≤i≤H, 1≤j≤W }, wherein H and W is respectively the height and width of original image, four subgraph difference
Are as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H'
Meet the following conditions respectively with W':
Further, when control module executes step Q2, the prediction of pixel in four subgraphs is calculated separately using fallout predictor
Error, prediction error are acquired by following formula:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is son
The prediction error of pixel in image.
Further, when control module executes step Q3, watermark insertion is successively carried out to four subgraphs according to prediction error
Processing, comprising the following steps:
Q31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to original image
Non-border pixel；
Q32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
The beneficial effects of the present invention are: the omnidirection context-prediction method and device towards reversible watermarking algorithm,
Although when watermark extracting, the pixel of front does not know the exact value of subsequent pixel value, water in reversible watermark of digital image
Printing extraction unit is not to know nothing to the information of pixel below, before watermark is extracted, though subsequent pixel value
It is so to be embedded in after watermark changed value, but since watermark is in order to guarantee invisibility, will not change with big amplitude
Pixel value, even value is also very close to original true value after insertion watermark, the pixel after these insertion watermarks is such as
Fruit is also contained in prediction context, then is conducive to fallout predictor and makes more accurate prediction.Therefore, the present invention passes through building base
In the fallout predictor of omnidirection context, original image is then decomposed into four subgraphs, calculates separately four using fallout predictor
Then the prediction error of pixel in subgraph successively carries out watermark insertion processing to four subgraphs according to prediction error, finally
It will be combined by four subgraphs of watermark insertion processing, form the final image for completing watermark insertion processing.Due to this
In the omnidirection context-prediction method of invention, other pixel predicted composition contexts around current pixel are utilized, and lead to
It crosses the prediction context to predict the value of current pixel, therefore can be greatly improved the precision of prediction.
Detailed description of the invention
The invention will be further described with example with reference to the accompanying drawing.
Fig. 1 is the flow chart of prediction error addition extended method of the invention；
Fig. 2 is the schematic diagram of the prediction context of fallout predictor.
Specific embodiment
- Fig. 2 referring to Fig.1, the omnidirection context-prediction method towards reversible watermarking algorithm, comprising the following steps:
S1, fallout predictor of the building based on omnidirection context；
S2, original image is decomposed into four subgraphs, calculates separately the pre- of pixel in four subgraphs using fallout predictor
Survey error；
S3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
S4, it will be combined by four subgraphs of watermark insertion processing, formed and complete the final of watermark insertion processing
Image.
Wherein, the fallout predictor based on omnidirection context, the formula of fallout predictor are constructed in step S1 are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in
The pixel of pixel x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor immediately below pixel x [i, j]
Pixel.
Wherein, original image is decomposed into four subgraphs in step S2, original image be I=x [i, j] | 1≤i≤H,
1≤j≤W }, wherein H and W is respectively the height and width of original image, and four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H'
Meet the following conditions respectively with W':
Wherein, calculate separately the prediction error of pixel in four subgraphs in step S2 using fallout predictor, prediction error by
Following formula acquires:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is son
The prediction error of pixel in image.
Wherein, watermark insertion processing, including following step are successively carried out to four subgraphs according to prediction error in step S3
It is rapid:
S31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to original image
Non-border pixel；
S32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
In addition, a kind of device for storing the omnidirection context-prediction method towards reversible watermarking algorithm, including control
Molding block and storage medium for storing control instruction, the control module read the control instruction in the storage medium simultaneously
Execute following steps:
Q1, fallout predictor of the building based on omnidirection context；
Q2, original image is decomposed into four subgraphs, calculates separately the pre- of pixel in four subgraphs using fallout predictor
Survey error；
Q3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
Q4, it will be combined by four subgraphs of watermark insertion processing, formed and complete the final of watermark insertion processing
Image.
Wherein, when control module executes step Q1, the fallout predictor based on omnidirection context, the formula of fallout predictor are constructed
Are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in
The pixel of pixel x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor immediately below pixel x [i, j]
Pixel.
Wherein, when control module executes step Q2, original image is decomposed into four subgraphs, original image is I={ x
[i, j] | 1≤i≤H, 1≤j≤W }, wherein H and W is respectively the height and width of original image, and four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H'
Meet the following conditions respectively with W':
Wherein, it when control module executes step Q2, is missed using the prediction that fallout predictor calculates separately pixel in four subgraphs
Difference, prediction error are acquired by following formula:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is son
The prediction error of pixel in image.
Wherein, when control module executes step Q3, successively four subgraphs are carried out at watermark insertion according to prediction error
Reason, comprising the following steps:
Q31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to original image
Non-border pixel；
Q32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
Specifically, although in reversible watermark of digital image, when watermark extracting, the pixel of front does not know subsequent pixel
The exact value of value, but watermark extracting unit is not to know nothing to the information of pixel below, before watermark is extracted,
Although subsequent pixel value is changed value after insertion watermark, but since watermark is in order to guarantee invisibility, will not
Pixel value is changed with big amplitude, even value is also very close to original true value after insertion watermark, these insertions
If the pixel after watermark is also contained in prediction context, is conducive to fallout predictor and makes more accurate prediction.Therefore, originally
Original image is then decomposed into four subgraphs, utilizes prediction by invention by fallout predictor of the building based on omnidirection context
Device calculates separately the prediction error of pixel in four subgraphs, then successively carries out watermark to four subgraphs according to prediction error
Insertion processing will finally be combined by four subgraphs of watermark insertion processing, be formed and complete watermark insertion processing most
Whole image.Since other pixel groups all around current pixel are utilized in omnidirection context-prediction method of the invention
At prediction context, and the value of current pixel is predicted by the prediction context, therefore can be greatly improved prediction
Precision.
Specifically, the fallout predictor in compression of images field is generally according to the relevant characteristic of neighborhood between pixel, with one by picture
The prediction context of other pixels composition around plain is reference, is predicted the value of current pixel, the phase between pixel
Closing property is all existing in all directions up and down, therefore predicts that the pixel in context is closer with current pixel and counts
When measuring more, predict more accurate.Since in compression of images, previously processed pixel is to the pixel value one of post-processing without institute
Know, therefore, Classical forecast device can only utilize the processed pixel in front, that is, the pixel of top and left, to form
Predict context, which limits the precision of predictions of Classical forecast device.However in reversible watermark of digital image, situation is then endless
Although complete in this way, in reversible watermark of digital image, when watermark extracting, the pixel of front does not know the accurate of subsequent pixel value
Value, but watermark extracting unit is not to know nothing to the information of pixel below, before watermark is extracted, subsequent picture
Although plain value is to be embedded in after watermark changed value, will not be with big width but since watermark is in order to guarantee invisibility
Degree change pixel value, even value is also very close to original true value after insertion watermark, after these insertion watermarks
If pixel is also contained in prediction context, is conducive to fallout predictor and makes more accurate prediction.According to this principle, originally
Invention proposes the fallout predictor based on omnidirection context, as shown in Fig. 2, Fig. 2 is the prediction based on omnidirection context
The schematic diagram of the prediction context of device, therefore, the formula of the fallout predictor are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in
The pixel of pixel x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor immediately below pixel x [i, j]
Pixel.
In order to guarantee that the invertibity of digital figure watermark, an important condition are sought in watermark extracting as each
Have passed through extension process pixel rebuild one with it when watermark is embedded in identical prediction context, that is, need to guarantee every
One pixel all uses identical prediction context when watermark is embedded in and is extracted.Therefore, omnidirection context of the invention is pre-
In survey method, original image is resolved into four subgraphs, watermark embedding operation successively then is carried out to this four subgraphs.Tool
Body, original image are I={ x [i, j] | 1≤i≤H, 1≤j≤W }, wherein H and W is respectively the height and width of original image
Degree, then four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H'
Meet the following conditions respectively with W':
In addition, the boundary pixel of original image is embedded in without watermark, but it is only involved in and forms omnidirectional prediction or more
Therefore text successively carries out watermark insertion processing to four subgraphs according to prediction error, comprising the following steps:
S31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to original image
Non-border pixel；
S32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
Specifically, when watermark is embedded in, first to U1It is handled, in this way for U1In each pixel u1[i, j], all
An omnidirectional prediction context can be constructed for it, the prediction context is by pixel x [2i-1,2j-1], x [2i-1,2j], x
[2i-1,2j+1], x [2i, 2j-1], x [2i, 2j+1], x [2i+1,2j-1], x [2i+1,2j] and x [2i+1,2j+1] composition.
Due to not handling the boundary pixel of original image, then x [2i, 2j-1] and x [2i, 2j+1] is from U2Pixel, x
[2i-1,2j] and x [2i+1,2j] are from U3Pixel, x [2i-1,2j-1], x [2i-1,2j+1], x [2i+1,2j-1] and x
[2i+1,2j+1] is from U4Pixel.U2、U3And U4All there are no be embedded in watermark, therefore U at this time1Prediction context in
All pixels are all original pixel values.
Then, then to U2Watermark insertion processing is carried out, for U2In each pixel u2[i, j], in omnidirectional prediction
Hereafter by x [2i-1,2j], x [2i-1,2j+1], x [2i-1,2j+2], x [2i, 2j], x [2i, 2j+2], x [2i+1,2j], x
[2i+1,2j+1] and x [2i+1,2j+2] composition.Due to not handling the boundary pixel of original image, then x [2i, 2j] and
X [2i, 2j+2] is from U1Pixel, x [2i-1,2j], x [2i-1,2j+2], x [2i+1,2j] and x [2i+1,2j+2] come
From U3Pixel, x [2i-1,2j+1] and x [2i+1,2j+1] are from U4Pixel.At this point, U1It has been embedded in watermark, therefore has been come
From U1Two pixels be embedded in watermark after pixel.
Then, then to U3Watermark insertion processing is carried out, for U3In each pixel u3[i, j], in omnidirectional prediction
Hereafter by x [2i, 2j-1], x [2i, 2j], x [2i, 2j+1], x [2i+1,2j-1], x [2i+1,2j+1], x [2i+2,2j-1], x
[2i+2,2j], x [2i+2,2j+1] composition.Due to not handling the boundary pixel of original image, then x [2i, 2j] and x
[2i+2,2j] is from U1Pixel, x [2i, 2j+1], x [2i+2,2j+1], x [2i, 2j-1] and x [2i+2,2j-1] come
From U2Pixel, x [2i+2,2j-1] and x [2i+2,2j+1] are from U4Pixel.At this point, U1And U2It has been embedded in watermark, because
This comes from U1And U2Four pixels be embedded in watermark after pixel.
Finally, to U4Watermark insertion processing is carried out, for U4In each pixel u4[i, j], above and below omnidirectional prediction
Text by x [2i, 2j], x [2i, 2j+1], x [2i, 2j+2], x [2i+1,2j], x [2i+1,2j+2], x [2i+2,2j], x [2i+2,
2j+1] and x [2i+2,2j+2] composition.Due to not handling the boundary pixel of original image, then x [2i, 2j], x [2i, 2j
+ 2], x [2i+1,2j+2] and x [2i+2,2j+1] is from U1Pixel, x [2i, 2j+1] and x [2i+2,2j+1] are from U2
Pixel, x [2i+1,2j] and x [2i+1,2j+2] are from U3Pixel.At this point, U1、U2And U3Watermark is all had been inserted into, therefore
Wherein eight all pixels are all the pixels being embedded in after watermark.
In watermark extracting, equally it is also required to resolving into final image into same four subgraph U1、U2、U3And U4, only
It is sequentially to need to come the reversed order of insertion in processing, that is, from U4、U3、U2To U1Sequence handled, this
Sample can guarantee in processing UiWhen, Ui+1To U4In pixel be all original pixel, and U1To Ui-1In pixel all be insertion
Pixel after watermark.This is identical as the situation in watermark telescopiny, that is, each pixel prediction context and its
Prediction context in watermark insertion is identical, in this way, it is ensured that original image can obtain nothing after watermark is extracted
The recovery of damage.
It is to be illustrated to preferable implementation of the invention, but the invention is not limited to above-mentioned embodiment party above
Formula, those skilled in the art can also make various equivalent variations on the premise of without prejudice to spirit of the invention or replace
It changes, these equivalent deformations or replacement are all included in the scope defined by the claims of the present application.
Claims (10)
1. the omnidirection context-prediction method towards reversible watermarking algorithm, it is characterised in that: the following steps are included:
S1, fallout predictor of the building based on omnidirection context；
S2, original image is decomposed into four subgraphs, is missed using the prediction that fallout predictor calculates separately pixel in four subgraphs
Difference；
S3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
S4, it will be combined by four subgraphs of watermark insertion processing, form the final image for completing watermark insertion processing.
2. the omnidirection context-prediction method according to claim 1 towards reversible watermarking algorithm, feature exist
In: the fallout predictor based on omnidirection context, the formula of the fallout predictor are constructed in the step S1 are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in pixel
The pixel of x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor the picture immediately below pixel x [i, j]
Element.
3. the omnidirection context-prediction method according to claim 2 towards reversible watermarking algorithm, feature exist
In: original image is decomposed into four subgraphs in the step S2, the original image be I=x [i, j] | 1≤i≤H, 1
≤ j≤W }, wherein H and W is respectively the height and width of original image, and four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H' and W'
Meet the following conditions respectively:
4. the omnidirection context-prediction method according to claim 3 towards reversible watermarking algorithm, feature exist
In: calculate separately the prediction error of pixel in four subgraphs in the step S2 using fallout predictor, the prediction error by with
Lower formula acquires:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is subgraph
In pixel prediction error.
5. the omnidirection context-prediction method according to claim 1 towards reversible watermarking algorithm, feature exist
In: watermark insertion processing is successively carried out to four subgraphs according to prediction error in the step S3, comprising the following steps:
S31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to the non-of original image
Boundary pixel；
S32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
6. a kind of device for storing the omnidirection context-prediction method towards reversible watermarking algorithm, it is characterised in that: packet
Control module and the storage medium for storing control instruction are included, the control that the control module is read in the storage medium refers to
It enables and executes following steps:
Q1, fallout predictor of the building based on omnidirection context；
Q2, original image is decomposed into four subgraphs, is missed using the prediction that fallout predictor calculates separately pixel in four subgraphs
Difference；
Q3, watermark insertion processing is successively carried out to four subgraphs according to prediction error；
Q4, it will be combined by four subgraphs of watermark insertion processing, form the final image for completing watermark insertion processing.
7. device according to claim 6, it is characterised in that: when the control module executes step Q1, building is based on complete
The fallout predictor of direction context, the formula of the fallout predictor are as follows:
Wherein,For the predicted value of pixel x [i, j], xnFor the pixel right above pixel x [i, j], xwFor in pixel
The pixel of x [i, j] front-left, xeFor the pixel in pixel x [i, j] front-right, xsFor the picture immediately below pixel x [i, j]
Element.
8. device according to claim 7, it is characterised in that: when the control module executes step Q2, original image
It is decomposed into four subgraphs, the original image is I={ x [i, j] | 1≤i≤H, 1≤j≤W }, wherein H and W is respectively original
The height and width of beginning image, four subgraphs are respectively as follows:
U1={ u1[i, j]=x [2i, 2j] | 1≤i≤H', 1≤j≤W'}
U2={ u2[i, j]=x [2i, 2j+1] | 1≤i≤H', 1≤j≤W'}
U3={ u3[i, j]=x [2i+1,2j] | 1≤i≤H', 1≤j≤W'}
U4={ u4[i, j]=x [2i+1,2j+1] | 1≤i≤H', 1≤j≤W'}
Wherein, H' and W' is respectively subgraph U1, subgraph U2, subgraph U3With subgraph U4Height and width, and H' and W'
Meet the following conditions respectively:
9. device according to claim 8, it is characterised in that: when the control module executes step Q2, utilize fallout predictor
The prediction error of pixel in four subgraphs is calculated separately, the prediction error is acquired by following formula:
Wherein, u [i, j] is the pixel in subgraph,For the predicted value of the pixel in subgraph, e [i, j] is subgraph
In pixel prediction error.
10. device according to claim 6, it is characterised in that: when the control module executes step Q3, missed according to prediction
Difference successively carries out watermark insertion processing to four subgraphs, comprising the following steps:
Q31, the pixel in four subgraphs is divided into respectively belongs to the boundary pixel of original image and belong to the non-of original image
Boundary pixel；
Q32, watermark insertion processing is successively carried out to the non-border pixel in four subgraphs according to prediction error.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810781894.0A CN109118417A (en) | 2018-07-16 | 2018-07-16 | Omnidirection context-prediction method and device towards reversible watermarking algorithm |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810781894.0A CN109118417A (en) | 2018-07-16 | 2018-07-16 | Omnidirection context-prediction method and device towards reversible watermarking algorithm |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN109118417A true CN109118417A (en) | 2019-01-01 |
Family
ID=64862711
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810781894.0A Withdrawn CN109118417A (en) | 2018-07-16 | 2018-07-16 | Omnidirection context-prediction method and device towards reversible watermarking algorithm |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109118417A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114399419A (en) * | 2021-12-16 | 2022-04-26 | 淮阴工学院 | Reversible image watermarking algorithm based on prediction error expansion |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110091092A1 (en) * | 2008-01-18 | 2011-04-21 | Mitek Systems | Systems for mobile image capture and remittance processing |
| CN102036079A (en) * | 2010-12-16 | 2011-04-27 | 北京航空航天大学 | Omni-directional prediction error histogram modification-based reversible image watermarking algorithm |
| CN105741221A (en) * | 2015-09-16 | 2016-07-06 | 湖南大学 | CAD engineering drawing reversible watermarking method, watermark embedding method and watermark extracting method |
-
2018
- 2018-07-16 CN CN201810781894.0A patent/CN109118417A/en not_active Withdrawn
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20110091092A1 (en) * | 2008-01-18 | 2011-04-21 | Mitek Systems | Systems for mobile image capture and remittance processing |
| CN102036079A (en) * | 2010-12-16 | 2011-04-27 | 北京航空航天大学 | Omni-directional prediction error histogram modification-based reversible image watermarking algorithm |
| CN105741221A (en) * | 2015-09-16 | 2016-07-06 | 湖南大学 | CAD engineering drawing reversible watermarking method, watermark embedding method and watermark extracting method |
Non-Patent Citations (3)
| Title |
|---|
| BO OUA, XIAOLONG LIB, YAO ZHAOA，RONGRONG NIA: "Reversible data hiding based on PDE predictor", 《THE JOURNAL OF SYSTEMS AND SOFTWARE》 * |
| 任克强；姚磊: "基于二次预测误差和像素排序的可逆图像水印", 《科学技术与工程》 * |
| 曾骁; 陈真勇; 陈明; 熊璋: "基于全方向预测与误差扩展的可逆数据隐藏", 《计算机研究与发展》 * |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114399419A (en) * | 2021-12-16 | 2022-04-26 | 淮阴工学院 | Reversible image watermarking algorithm based on prediction error expansion |
| CN114399419B (en) * | 2021-12-16 | 2024-03-26 | 淮阴工学院 | Reversible image watermarking algorithm based on prediction error expansion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107844779B (en) | Video key frame extraction method | |
| CN101651837B (en) | Reversible video frequency watermark method based on interframe forecast error histogram modification | |
| CN112365404B (en) | Contact net panoramic image splicing method, system and equipment based on multiple cameras | |
| CN104952083B (en) | A kind of saliency detection method based on the modeling of conspicuousness target background | |
| CN106023058A (en) | High-fidelity reversible watermark embedding method | |
| US20170094306A1 (en) | Method of acquiring neighboring disparity vectors for multi-texture and multi-depth video | |
| Yang et al. | Efficient reversible data hiding algorithm based on gradient-based edge direction prediction | |
| CN109977895A (en) | A kind of wild animal video object detection method based on multi-characteristic fusion | |
| CN109344820A (en) | Digital electric meter Recognition of Reading method based on computer vision and deep learning | |
| CN102236887A (en) | Motion-blurred image restoration method based on rotary difference and weighted total variation | |
| CN105741225A (en) | Reversible watermark method of multi-dimensional prediction error extension | |
| CN109360140B (en) | Reversible image watermarking method and device based on prediction error addition expansion | |
| CN103617591B (en) | A kind of dual-color image blind watermarking method decomposed based on QR and compensate | |
| CN110009621A (en) | One kind distorting video detecting method, device, equipment and readable storage medium storing program for executing | |
| CN106485640A (en) | A kind of reversible water mark computational methods based on multi-level IPVO | |
| CN109118417A (en) | Omnidirection context-prediction method and device towards reversible watermarking algorithm | |
| CN102117474B (en) | Digital picture watermark embedding and detecting method and device | |
| CN109859090B (en) | Reversible watermarking method and device based on human visual system | |
| CN112862692A (en) | Image splicing method applied to underground coal mine roadway | |
| CN106603888A (en) | Image color extraction processing structure | |
| CN103440616A (en) | High volume reversible watermarking method based on self-adaptive prediction model | |
| CN101056411B (en) | Method for detecting the image displacement | |
| CN108200435B (en) | Reversible information hiding method for fractal image | |
| CN101916432B (en) | Parallelization method for image segmentation algorithm based on mean shift | |
| CN110248182B (en) | Scene segment shot detection method |
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 | ||
| WW01 | Invention patent application withdrawn after publication | ||
| WW01 | Invention patent application withdrawn after publication |
Application publication date: 20190101 |