CN113837915B - Ceramic watermark model training method and embedding method for binaryzation of boundary region - Google Patents
Ceramic watermark model training method and embedding method for binaryzation of boundary region Download PDFInfo
- Publication number
- CN113837915B CN113837915B CN202110844701.3A CN202110844701A CN113837915B CN 113837915 B CN113837915 B CN 113837915B CN 202110844701 A CN202110844701 A CN 202110844701A CN 113837915 B CN113837915 B CN 113837915B
- Authority
- CN
- China
- Prior art keywords
- image
- training
- ceramic
- watermark
- pixel selection
- 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.)
- Active
Links
- 239000000919 ceramic Substances 0.000 title claims abstract description 231
- 238000012549 training Methods 0.000 title claims abstract description 180
- 238000000034 method Methods 0.000 title claims abstract description 96
- 238000005260 corrosion Methods 0.000 claims description 31
- 230000007797 corrosion Effects 0.000 claims description 31
- 238000010304 firing Methods 0.000 claims description 29
- 238000012545 processing Methods 0.000 claims description 26
- 230000003628 erosive effect Effects 0.000 claims description 21
- 238000010586 diagram Methods 0.000 claims description 19
- 230000006835 compression Effects 0.000 claims description 14
- 238000007906 compression Methods 0.000 claims description 14
- 238000009826 distribution Methods 0.000 claims description 6
- 238000002347 injection Methods 0.000 claims description 6
- 239000007924 injection Substances 0.000 claims description 6
- 238000012546 transfer Methods 0.000 claims description 6
- 238000007781 pre-processing Methods 0.000 claims description 5
- 230000008569 process Effects 0.000 abstract description 35
- 238000007650 screen-printing Methods 0.000 abstract description 15
- 238000005516 engineering process Methods 0.000 abstract description 13
- 238000000605 extraction Methods 0.000 abstract description 12
- 230000006870 function Effects 0.000 description 77
- 230000015654 memory Effects 0.000 description 14
- 238000004422 calculation algorithm Methods 0.000 description 10
- 238000007639 printing Methods 0.000 description 9
- 238000003860 storage Methods 0.000 description 8
- 238000005070 sampling Methods 0.000 description 7
- 239000003086 colorant Substances 0.000 description 6
- 238000013461 design Methods 0.000 description 6
- 230000000694 effects Effects 0.000 description 6
- 230000010339 dilation Effects 0.000 description 5
- 239000000049 pigment Substances 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 4
- 230000009466 transformation Effects 0.000 description 4
- 239000000969 carrier Substances 0.000 description 3
- 238000006243 chemical reaction Methods 0.000 description 3
- 238000004590 computer program Methods 0.000 description 3
- 238000007641 inkjet printing Methods 0.000 description 3
- 238000012986 modification Methods 0.000 description 3
- 230000004048 modification Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 238000004364 calculation method Methods 0.000 description 2
- 230000008859 change Effects 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 238000005530 etching Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000007246 mechanism Effects 0.000 description 2
- 238000002156 mixing Methods 0.000 description 2
- 239000007787 solid Substances 0.000 description 2
- 239000000243 solution Substances 0.000 description 2
- 230000000007 visual effect Effects 0.000 description 2
- 238000003491 array Methods 0.000 description 1
- 238000005467 ceramic manufacturing process Methods 0.000 description 1
- 239000013043 chemical agent Substances 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000010422 painting Methods 0.000 description 1
- 229910052573 porcelain Inorganic materials 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000012216 screening Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
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
- G06T1/005—Robust watermarking, e.g. average attack or collusion attack resistant
-
- C—CHEMISTRY; METALLURGY
- C04—CEMENTS; CONCRETE; ARTIFICIAL STONE; CERAMICS; REFRACTORIES
- C04B—LIME, MAGNESIA; SLAG; CEMENTS; COMPOSITIONS THEREOF, e.g. MORTARS, CONCRETE OR LIKE BUILDING MATERIALS; ARTIFICIAL STONE; CERAMICS; REFRACTORIES; TREATMENT OF NATURAL STONE
- C04B35/00—Shaped ceramic products characterised by their composition; Ceramics compositions; Processing powders of inorganic compounds preparatory to the manufacturing of ceramic products
- C04B35/622—Forming processes; Processing powders of inorganic compounds preparatory to the manufacturing of ceramic products
- C04B35/64—Burning or sintering processes
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/10—Protecting distributed programs or content, e.g. vending or licensing of copyrighted material ; Digital rights management [DRM]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/20—Image enhancement or restoration using local operators
- G06T5/30—Erosion or dilatation, e.g. thinning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/70—Denoising; Smoothing
-
- 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
- G06T2201/0065—Extraction of an embedded watermark; Reliable detection
-
- 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
- G06T2207/10—Image acquisition modality
- G06T2207/10024—Color image
-
- 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
- G06T2207/20081—Training; Learning
-
- 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
- G06T2207/20084—Artificial neural networks [ANN]
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Software Systems (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Biomedical Technology (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- Biophysics (AREA)
- Health & Medical Sciences (AREA)
- Mathematical Physics (AREA)
- Artificial Intelligence (AREA)
- Life Sciences & Earth Sciences (AREA)
- Data Mining & Analysis (AREA)
- Ceramic Engineering (AREA)
- Manufacturing & Machinery (AREA)
- Multimedia (AREA)
- Computer Security & Cryptography (AREA)
- Inorganic Chemistry (AREA)
- Computer Hardware Design (AREA)
- Technology Law (AREA)
- Materials Engineering (AREA)
- Structural Engineering (AREA)
- Organic Chemistry (AREA)
- Editing Of Facsimile Originals (AREA)
- Image Processing (AREA)
Abstract
The invention discloses a ceramic watermark model training method and an embedding method for binaryzation of a boundary region, which can ensure that an encoder can embed secret information into the boundary region of a ceramic trademark in an invisible manner by means of a binaryzation technology, ensure that the boundary color only presents the background color of the ceramic trademark or the color of the trademark, further reduce the difficulty of manual color matching of a technologist, solve the problem of inaccurate watermark extraction caused by a screen printing process and improve the extraction accuracy of copyright information.
Description
Technical Field
The invention relates to the technical field of ceramics, in particular to a ceramic watermark model training method and an embedding method for binaryzation of a boundary region.
Background
In recent years, with the development of digital watermarking technology, the application field of digital watermarking is also expanded, and especially, huge achievements are obtained on copyright protection, hidden identification, authentication and secure invisible communication, however, for ceramic special carriers, the digital watermarking technology is based on digital carriers of images, videos, audios and the like, and the ceramic carriers are required to be fired through various traditional processes, so that the digital watermarking technology cannot be directly grafted on ceramics.
At present, a ceramic copyright authentication network design based on a bottom changing mechanism exists, watermarks can be embedded in the boundaries of ceramic trademarks in an imperceptible mode through the network, and the problem of inaccurate watermark extraction caused by a screen printing process is solved.
However, in practice it has been found that when embedding secret information using the above described scheme, there is often a pronounced gradient of colour on the ceramic trademark border, as shown in fig. 1, the colour of the ceramic trademark border appears as grey and grey as a gradient. For the screen printing process, each color in the ceramic trademark is manually configured, and in the face of the "gradient" effect, it is difficult for a technologist to accurately match colors and express one by one. Generally, the types of colors that can be manually configured are very limited, and it is difficult to configure corresponding gradient color patterns, which may cause the phenomenon of decoding failure due to the fact that the ceramic trademark cannot be expressed by screen printing due to pixel modification.
Disclosure of Invention
In view of this, the embodiment of the present invention provides a ceramic watermark model training method and an embedding method for binarization of a boundary region, so as to solve the problem that an obvious gradient color appears on the boundary of a ceramic trademark.
According to a first aspect, an embodiment of the present invention provides a ceramic watermark model training method for boundary region binarization, including:
respectively acquiring training images and training watermark information;
carrying out corrosion operation on the training image to obtain a training corrosion image, and carrying out expansion operation on the training image to obtain a training expansion image;
inputting the training image and the training watermark information into a current encoder to generate a binary pixel selection graph, and obtaining a watermark image according to the binary pixel selection graph, the training erosion image and the training expansion image;
calculating the current loss value of the binary pixel selection graph loss function, and adjusting the current encoder according to the current loss value of the binary pixel selection graph loss function to obtain an updated current encoder;
putting the watermark image into a preset noise layer for noise processing;
sending the watermark image subjected to noise processing into a current decoder for decoding to obtain secret information, obtaining a loss value of a cross entropy loss function according to the secret information and the training watermark information, and updating the current decoder according to the loss value of the cross entropy loss function to obtain an updated current decoder.
The ceramic watermark model training method for boundary area binarization provided by the embodiment of the invention has the advantages that by means of the binarization technology, the encoder can not only embed secret information into the boundary area of the ceramic trademark in an invisible manner, but also ensure that the boundary color only presents the background color of the ceramic trademark or the color of the trademark, further reduce the difficulty of manual color matching of a technologist, solve the problem of inaccurate watermark extraction caused by a screen printing process, and improve the extraction accuracy of copyright information.
With reference to the first aspect, in a first implementation manner of the first aspect, obtaining a watermark image according to the binary pixel selection map, the training erosion image, and the training dilation image includes: performing dot multiplication on the binary pixel selection image and the training corrosion image to obtain a corrosion selection image; performing reverse color processing on the binary pixel selection image, and performing dot multiplication on the binary pixel selection image subjected to the reverse color processing and the training expansion image to obtain an expansion selection image; and superposing the corrosion selection diagram and the expansion selection diagram to obtain the watermark image.
With reference to the first aspect or the first implementation manner of the first aspect, in a second implementation manner of the first aspect, before adjusting the current encoder according to the binary pixel selection map loss function loss value, the method further includes: obtaining a weight value of a loss function of the binary pixel selection graph; adjusting the current encoder according to the binary pixel selection map loss function loss value comprises: updating the current encoder by using the weight value of the binary pixel selection graph loss function and the loss value of the binary pixel selection graph loss function; before updating the current decoder according to the cross entropy loss function loss value, the method further includes: acquiring a weight value of a cross entropy loss function; updating the current decoder according to the cross-entropy loss function loss value comprises: and updating the current decoder by using the weight value of the cross entropy loss function and the loss value of the cross entropy loss function.
With reference to the second embodiment of the first aspect, in a third embodiment of the first aspect, before the preset step number threshold, the weight value of the binary pixel selection map loss function is 0; and after the step number threshold value, setting the weight value of the loss function of the binary pixel selection graph to be more than 0 and less than the weight value of the cross entropy loss function.
With reference to the third implementation manner of the first aspect, in the fourth implementation manner of the first aspect, before the obtaining the training image and the training watermark information respectively, the method further includes: and carrying out first numerical clipping on the loss function value of the binary pixel selection map.
With reference to the fourth implementation manner of the first aspect, in the fifth implementation manner of the first aspect, after the calculating a current loss value of the binary pixel selection map loss function, the method further includes: judging whether the pixel distribution in the binary pixel selection image reaches a preset condition or not according to the current loss value of the binary pixel selection image loss function; and performing second numerical clipping on the loss function value of the binary pixel selection image after the pixel distribution in the binary pixel selection image reaches a preset condition.
According to a second aspect, an embodiment of the present invention provides a ceramic watermark model training apparatus with binarized boundary regions, including:
the acquisition module is used for respectively acquiring training images and training watermark information;
the preprocessing module is used for carrying out corrosion operation on the training image to obtain a training corrosion image and carrying out expansion operation on the training image to obtain a training expansion image;
the watermark generating module is used for inputting the training image and the training watermark information into the current encoder to generate a binary pixel selection image, and obtaining a watermark image according to the binary pixel selection image, the training erosion image and the training expansion image;
the first adjusting module is used for calculating a loss value of a binary pixel selection map loss function, and adjusting the current encoder according to the loss value of the binary pixel selection map loss function to obtain an updated current encoder;
the noise processing module is used for putting the watermark image into a preset noise layer for noise processing;
and the second adjusting module is used for sending the watermark image subjected to noise processing into a current decoder for decoding to obtain secret information, obtaining a loss value of a cross entropy loss function according to the secret information and the training watermark information, and updating the current decoder according to the loss value of the cross entropy loss function to obtain an updated current decoder.
With reference to the second aspect, in a first embodiment of the second aspect, the noise layer includes: geometric distortion, motion blur, color shift, gaussian noise, and JEPG compression.
With reference to the first embodiment of the second aspect, in the second embodiment of the second aspect, the distortion coefficient of the geometric distortion is less than 1; and/or, the motion blur adopts a linear blur kernel, the pixel width of the linear kernel is not more than 10, the linear angle is randomly selected, and the range is not more than 1/2 pi; and/or the offset value of the color offset is required to be uniformly distributed, and the offset value is-0.2-0.3; and/or, the compression quality factor of the JEPG compression is greater than 50.
According to a third aspect, an embodiment of the present invention further provides an encoder, which is obtained by training with the boundary region binarization ceramic watermark model training method described in the first aspect or any implementation manner of the first aspect.
According to a fourth aspect, an embodiment of the present invention further provides a decoder, which is obtained by training with the boundary region binarization ceramic watermark model training method described in the first aspect or any implementation manner of the first aspect.
According to a fifth aspect, an embodiment of the present invention further provides a ceramic watermark embedding method for boundary region binarization, including: respectively acquiring training images and training watermark information; carrying out corrosion operation on the training image to obtain a corrosion image, and carrying out expansion operation on the training image to obtain an expansion image; inputting the training image and the training watermark information into the encoder of the third aspect for encoding to obtain a binary pixel selection map, and obtaining an electronic watermark image by using the binary pixel selection map, the erosion selection map and the expansion selection map; and after the electronic watermark image is transferred to the ceramic prefabricated product, firing the ceramic prefabricated product to obtain the ceramic with the watermark image.
With reference to the fifth aspect, in a first embodiment of the fifth aspect, the transferring the electronic watermark image onto the ceramic preform includes: inputting the electronic watermark image into a preset ceramic ink-jet injection machine, and carrying out ink jet on the ceramic prefabricated product by using the ceramic ink-jet injection machine so as to transfer the electronic watermark image onto the ceramic prefabricated product; or generating paper edition stained paper according to the electronic watermark image; and (3) paving the paper pattern paper on the ceramic prefabricated product to transfer the electronic watermark image to the ceramic prefabricated product.
With reference to the fifth aspect, in a second embodiment of the fifth aspect, when the ceramic preform is a commodity ceramic preform, firing the ceramic preform to obtain a ceramic with a watermark pattern comprises: firing the daily ceramic prefabricated product at 800-1380 ℃ to obtain daily ceramic; when the ceramic preform is a sanitary ceramic preform, firing the ceramic preform to obtain a ceramic having a watermark pattern comprises: firing the sanitary ceramic prefabricated product at 800-1380 ℃ to obtain sanitary ceramic; when the ceramic preform is an architectural ceramic preform, firing the ceramic preform to obtain a ceramic with a watermark pattern comprising: and firing the architectural ceramic prefabricated product at 800-1380 ℃ to obtain the architectural ceramic.
According to a sixth aspect, an embodiment of the present invention further provides a decryption method for a ceramic watermark binarized in a boundary region, including: positioning the watermark pattern on the ceramic; inputting the positioned watermark pattern into the decoder of the fourth aspect for decoding to obtain the watermark information in the watermark pattern.
Drawings
The features and advantages of the present invention will be more clearly understood by reference to the accompanying drawings, which are illustrative and not to be construed as limiting the invention in any way, and in which:
FIG. 1 is a schematic view of a gradient color on the border of a ceramic trademark;
FIG. 2 is a schematic diagram of binarization at the boundary of a ceramic trademark;
FIG. 3 is a frame diagram of a ceramic copyright network based on a binarization technique;
FIG. 4 is a schematic view of an encoder model;
FIG. 5 is a schematic diagram of a decoder model;
FIG. 6 is a schematic view of a ceramic label;
FIG. 7 is a binary pixel selection map corresponding to the ceramic trademark diagram of FIG. 6;
FIG. 8 is a graph showing the results after the etching operation;
FIG. 9 is a graph showing the results after the expansion operation;
fig. 10 is a schematic flowchart of a ceramic watermark model training method based on boundary region binarization in embodiment 1 of the present invention;
fig. 11 is a schematic structural diagram of a ceramic watermark model training apparatus with boundary region binarization in embodiment 2 of the present invention;
FIG. 12 is a flow chart of a method for making a ceramic watermark pattern based on an ink jet process;
FIG. 13 is a flow chart of a method of making a ceramic watermark pattern based on screen printing;
FIG. 14 is a schematic flow chart of ceramic copyright encryption and decryption based on the ink-jet process;
fig. 15 is a schematic flow chart of ceramic copyright encryption and decryption based on screen printing.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, 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 some, but not all, embodiments of the present invention. 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.
Example 1
In order to solve the problem that an obvious gradient color appears on the boundary of the ceramic trademark, embodiment 1 of the invention provides a ceramic watermark model training method with binaryzation in a boundary area. By means of the binarization technology, the encoder can embed the secret information into the boundary area of the ceramic trademark in an invisible mode, and can ensure that the boundary color only presents the background color of the ceramic trademark or the color of the trademark, as shown in fig. 2, the boundary color of the ceramic trademark presents black or white, namely binarization. Through the binarization network algorithm provided by the embodiment of the invention, the phenomenon of gradual color change can be effectively eliminated, the difficulty of manual color matching of a technologist is further reduced, the problem of inaccurate watermark extraction caused by a screen printing process is solved, and the extraction accuracy of copyright information is improved.
A network frame diagram of a ceramic watermark model with binarized boundary regions is shown in fig. 3, and comprises a plurality of network sub-modules such as an encoder, a decoder, a noise layer network, a binarizing network (hereinafter referred to as a binarizing network) composed of a binarizing technology algorithm, and the like. The encoder embeds the secret information into the carrier image to generate a secret image (hereinafter referred to as watermark image) containing watermark information. The decoder extracts the secret information embedded by the watermark image. The noise layer network simulates noise attack to the watermark image in the printing, shooting and ceramic firing processes, wherein the noise attack comprises perspective transformation attack, brightness noise attack, saturation noise attack, color saturation noise attack, Gaussian noise attack and Jpeg compression noise attack. The binarization network can ensure that the current encoder embeds the secret information on the boundary of the carrier image in an invisible mode and ensure that the color of the boundary area presents a binarization effect.
a) The encoder model comprises a down-sampling convolution layer and an up-sampling convolution part as shown in fig. 4, wherein the down-sampling convolution part is used for performing convolution calculation on a carrier image to perform feature extraction to form a feature body with high dimensionality, the up-sampling convolution part is used for performing feature addition on the feature body of each layer of the down-sampling convolution and an up-sampled input to gradually restore image details, and finally a pixel selection graph with the number of channels corresponding to the carrier image being 1 is generated. Here, the pixel selection diagram (single-channel residual image) in the present application is a binary pixel selection image, which will be explained in detail later. In addition, it should be noted that the pixel selection map is different from the conventional 3-channel residual map, and the pixel selection map can be directly overlapped with the original 3-channel RGB color image.
b) The decoder model is shown in fig. 5, which contains a downsampled convolution module and a full connection layer module. The down-sampling module is used for performing convolution calculation on the watermark image, extracting watermark features to form a watermark information feature map, and the full-connection layer further compresses the watermark information feature map and converts the watermark information feature map into a binary bit sequence so as to realize secret information extraction.
c) The noise layer is used for simulating various physical attacks which may be suffered in printing shooting, and comprises perspective transformation attack, Gaussian noise attack, color shift, motion blur, brightness contrast transformation and JEPG compression attack, and attack such as ceramic firing and the like. It should be noted that all the attacks are random values, and the value range setting changes according to the environmental change.
d) Aiming at the phenomenon of 'gradient color' generated by a ceramic process, the embodiment 1 of the invention provides a binary network design algorithm, which is also the core content of the algorithm of the invention, and the following description is provided:
d _1) introduction of elements formed by the binarization algorithm, mainly comprising A, B, residual and residual _ loss.
As defined herein: a is an image generated by an original trademark graph through corrosion operation, B is an image generated by the original trademark graph through dilation operation, and residual refers to a binary pixel selection graph (namely, a residual graph proposed in the step a) generated through iterative training of an encoder network. residual is a single-channel binary image (the number of channels is 1) with the same size as a and B, and the function of the residual is to select corresponding pixels in a or B for output according to needs, for example, black (the pixel value is 0) indicates that the pixels in the image a are selected for output, and white (the pixel value is 1) indicates that the pixels in the image B are selected for outputOutput, for example, fig. 6 is a ceramic trademark diagram, and fig. 7 is a binary pixel selection diagram corresponding to the ceramic trademark diagram of fig. 6. Loss residual And a distortion function used for guiding network training, namely a binary pixel selection graph loss function.
d _2) principle and idea for realizing binary image
I) Generation and schematic of A and B
After the etching operation, the edge area of the a image shrinks inward as shown in fig. 8, and the edge pixel width becomes smaller, so that the edge becomes thinner compared to the original image. After the dilation operation, the edge area of the B image is expanded outward as shown in fig. 9, and the edge pixel width is increased, so that the edge is thicker than that of the original image.
II) Generation of a binary map
If a pixel selection image residual containing secret information is trained through the network, the dense binary output image can be expressed by the following formula, that is, the image does not contain the gradient.
I' ═ a residual + B (1-residual) formula 1
Wherein, I' represents the dense ideal binary output image. If the residual corresponding value is 1, that is, residual is 1, (1-residual) is 0, then a residual is used to indicate that the corresponding pixel in a is extracted and assigned to I'; and B (1-residual) is set to zero. Conversely, if the residual corresponding value is 0, then residual is 0, (1-residual) is 1, and the corresponding pixel in B is assigned to I'.
As shown in fig. 8 and 9, considering that a and B correspond to the eroded and expanded versions of the same original map, the difference is completely consistent only in the boundary portion and in the flat background portion. Based on this, it can be considered that the output image I' containing dense binarization essentially selects an original boundary color (i.e. color in a or B) by 0/1 modulation in residual on the basis of keeping the original image background. Therefore, the real binary image output can be realized.
In addition, it is necessary to explain: the binarization of the output image needs to be guaranteed by binarization of residual. If the numerical value in residual is not 0/1, a new gradient color will appear in I'. Based on this, the way how to guarantee that residual tends to value 0/1 in training is given later.
d _3) training and binarization guarantee of residual
In embodiment 1 of the invention, residual is obtained by network iterative training through a deep network, and at the initial stage of network training, the pixel values of the residual image are randomly distributed in the network. Therefore, the invention needs to normalize the pixel value in residual to 0, 1 in each training process during training]In between, and in order to realize residual binarization (pixel value is either 0 or 1 in the image), the invention designs Loss residual The network training is guided by a distortion function (also called a binary pixel selection graph loss function), and the specific formula is as follows:
e) in order to ensure the above-mentioned cooperative training of multiple network modules, other conventional network distortions are designed as follows.
Total Loss function by Loss serect Loss function and Loss residual Loss function composition, said Loss serect The loss function is a cross entropy loss function, X and Y respectively represent a watermark sequence input by the coding network and a watermark sequence output by the decoding network, X is represented by a binary sequence of 0 and 1, and Y is represented by a probability between 0 and 1 as follows:
the Loss residual The function of the binary pixel selection map loss function is to optimize the pixel value of the binary pixel selection map loss function so that the pixel value is either 0 or 1. Where residual represents the residual map generated by the current encoder, and its channel is 1, which is a binary map corresponding to the original trademark.
In order to better adjust the network model and control the balance relationship between the visual quality of the watermark image and the robustness of the watermark extraction network, corresponding weight values are respectively added to all the loss functions. W serect ，W residual Respectively, the weight values of the corresponding loss functions, and the total loss function is expressed as follows:
Loss＝W serect Loss serect +W residual Loss residual equation 4
Based on this, embodiment 1 of the present invention provides a training method for a ceramic watermark model with binarized boundary region, fig. 10 is a schematic flow chart of the training method for a ceramic watermark model with binarized boundary region in embodiment 1 of the present invention, as shown in fig. 10, the training method for a ceramic watermark model with binarized boundary region in embodiment 1 of the present invention includes the following steps:
s101: and respectively acquiring training images and training watermark information.
In embodiment 1 of the present invention, an LLD-Logo in a Large Logo Dataset (LLD) Dataset of a training set is first prepared, wherein symbol images (i.e., training images) with different resolutions of 64 × 64 to 400 × 400 are included. The LLD data set was preprocessed and scaled to 256 × 256 resolution.
In embodiment 1 of the present invention, the watermark information may be a binary watermark sequence. The binary watermark sequence may be appropriately deformed to have the same size as any of the training images in the preprocessed LLD data set in step S101.
S102: and carrying out corrosion operation on the training image to obtain a training corrosion image, and carrying out expansion operation on the training image to obtain a training expansion image.
In embodiment 1 of the present invention, the training erosion image is an image generated by the training image through erosion operation, and after the training image is subjected to erosion operation, an edge area of the training image is shrunk inward, and a width of an edge pixel is reduced, so that an edge of the training image is thinner than that of the training image. The training dilated image is an image generated by the training image through a dilation operation. The training image is expanded outward in the edge area after the expansion operation, and the width of the edge pixel is increased, so that the edge is thicker than that of the training image.
S103: inputting the training image and the training watermark information into a current encoder to generate a binary pixel selection graph, and obtaining a watermark image according to the binary pixel selection graph, the training erosion image and the training expansion image.
As a specific implementation manner, the following technical scheme may be adopted to obtain the watermark image according to the binary pixel selection map, the training erosion image, and the training dilation image: performing dot multiplication on the binary pixel selection image and the training corrosion image to obtain a corrosion selection image; performing reverse color processing on the binary pixel selection image, and performing dot multiplication on the binary pixel selection image subjected to the reverse color processing and the training expansion image to obtain an expansion selection image; and superposing the corrosion selection diagram and the expansion selection diagram to obtain the watermark image.
For example, the watermark image may be obtained by using the above formula 1, I '═ residual + B (1-residual), where I' may also be referred to as the watermark image, a residual is the erosion selective map, and B (1-residual) is the expansion selective map. For example, when a × residual, the value of pixel 1 in residual is dot-multiplied by a, and then the result of a × residual is that the value of pixel 1 corresponding to residual will be left, and the value of pixel 0 corresponding to residual will be discarded. Then (1-residual) B, since the residual is inverted, the truncated value in case one will be offset by the pixel in the corresponding position of the B map.
S104: and calculating the current loss value of the binary pixel selection image loss function, and adjusting the current encoder according to the current loss value of the binary pixel selection image loss function to obtain the updated current encoder.
In embodiment 1 of the present invention, the Loss function of the binary pixel selection map is the Loss function of the Loss residual A loss function.
As a specific implementation manner, before adjusting the current encoder according to the loss value of the binary pixel selection map loss function, the method further includes: and acquiring the weight value of the loss function of the binary pixel selection map. Further, adjusting the current encoder according to the loss value of the binary pixel selection map loss function includes: and updating the current encoder by using the weight value of the binary pixel selection graph loss function and the loss value of the binary pixel selection graph loss function. Before a preset step threshold value, the weight value of the binary pixel selection graph loss function is 0; after the step number threshold, a weight value of the cross entropy loss function is greater than a weight value of the binary pixel selection map loss function.
As a further embodiment, before calculating the binary pixel selection map loss function loss value between the watermark image and the training image, the method further includes: and performing numerical clipping on the binary pixel selection map loss function before the step number threshold.
That is to say, in the initial stage of training, the pixel values in the residual image are subjected to first gradient clipping, the clipping range is (-M, M), and the value range of M is-0.5 to 0.5.
As a further embodiment, after the calculating the current loss value of the binary pixel selection map loss function, the method further includes: judging whether the pixel distribution in the binary pixel selection image reaches a preset condition or not according to the current loss value of the binary pixel selection image loss function; and after the pixel distribution in the binary pixel selection image reaches a preset condition, performing second numerical clipping on the loss function value of the binary pixel selection image.
S105: and putting the watermark image into a preset noise layer for noise processing.
In embodiment 1 of the present invention, in order to make the watermark image able to resist distortion in the printing or shooting process, a noise layer capable of simulating a real physical scene is designed between the current encoder and the current decoder, so as to simulate various noises that may exist in the watermark image in the ceramic manufacturing process. When the copyright watermark information is embedded into the current encoder, the visual consistency of the output watermark pattern and the original input pattern needs to be ensured as much as possible so as to ensure the final ceramic presentation effect.
Based on the mechanism, the counter-generation type digital watermark model can generate a robust watermark image which can resist ceramic making attack on one hand, and on the other hand, the invisibility of the image vision of the watermark after being embedded is ensured. In order to ensure the concrete implementation of this technology, the following focuses on a noise floor design concept that is resistant to ceramic processes.
The noise floor design of the ceramic copyright certification technology process based on screen printing is described below.
The ceramic stained paper is special ceramic stained paper printed on the surface of ceramic (or porcelain blank), and the manufacturing process comprises the following steps:
step 1: and (3) making a stained paper, wherein the making is to convert the provided ceramic pattern into an AI file required for making the stained paper.
Step 2: the plate burning is a film for making trademarks or patterns on the flower surface, and is similar to a negative film of a camera.
And step 3: the color matching is to combine various primary colors of the ceramic pigment according to a certain proportion to provide the color required by the ceramic trademark.
And 4, step 4: and (4) sampling, namely putting the adjusted color pigment and the manufactured printing plate into a semi-automatic pattern paper machine to form the pattern paper.
In the process of transferring the ceramic watermark pattern to the ceramic, each procedure generates noise attack and has important influence on correctly extracting watermark information by a decoding network, so that the noise attack caused by each procedure needs to be simulated, and the specific description is as follows: in step 1, the ceramic watermark pattern is subjected to a JEPG compression operation when the corresponding AI document is manufactured. In step 2, the ceramic watermark pattern needs to be exposed by chemical agents when passing through the printing process, and the step has certain influence on the brightness, the contrast, the color and the tone of the ceramic watermark pattern. In step 3, the toning is divided into manual toning and machine toning. When the colors in the ceramic watermark pattern exceed four colors, manual color matching is needed, and the manual color matching can cause color deviation of the ceramic pattern. The resulting color shift is negligible because the machine toning is precise. Based on the analysis, the invention builds a noise layer network capable of simulating all process attacks, wherein the noise layer network comprises geometric distortion, motion blur, color shift, Gaussian noise and JEPG compression. The motion blur and the geometric distortion are mainly used for simulating noise attack for shooting ceramic watermark patterns to carry out copyright authentication. The five attack noises are randomly valued in a certain range, and the noise attack in the process of transferring the ceramic watermark image electronic plate into the paper plate is fully simulated. The noise layer designed aiming at the screen printing process is combined with the counter-generating digital watermark model algorithm, so that the feasibility of the ceramic watermark authentication framework based on the screen printing process is ensured.
The following mainly describes the noise layer design of the ceramic copyright certification technology process based on ink jet printing. The ink-jet technology is essentially that a ceramic watermark image is pre-stored in an automatic ink-jet computer, the computer carries out color matching according to the ceramic watermark image, and then the ink-jet computer carries out painting on a ceramic carrier. The ink jet printer may cause certain color errors when color matching is performed, which may have certain influence on the color and tone of the ceramic watermark pattern. Furthermore, since the color pigments are drawn directly onto the ceramic support, the effect of the ceramic support material itself on the pigments, including brightness, contrast, color and hue, cannot be neglected. Furthermore, since the verification stage of copyright information follows, geometric distortion and motion blur also need to be considered. Based on this, the noise floor attack against the inkjet process is mainly: geometric distortion, motion blur, color shift, and gaussian noise. The four attack noises are randomly valued in a certain range, and the noise attack of the ceramic watermark image drawn on the ceramic carrier is fully simulated. The noise layer designed aiming at the ink jet printing process is combined with the counter-generating digital watermarking algorithm, so that the feasibility of the ceramic watermarking authentication framework based on the ink jet printing process is ensured.
Because the ceramic trademark printing process is under the high temperature condition of 700-1100 ℃, the pigment colored by the ceramic is greatly influenced by temperature, humidity and air atmosphere, so that the image distortion is also large, and the watermark image contrast, saturation and color tone distortion range are larger. In addition, when the ceramic screen printing is carried out, the color offset phenomenon can also occur to a certain extent when the printing ink is influenced by the temperature, and the watermark image can still extract the watermark information without distortion after the image is distorted, so that the network constructs a noise layer between the current encoder and the current decoder. The noise layer is constructed mainly for simulating the attack situation possibly suffered by the ceramic firing process, namely, the distortion possibly caused by the ceramic printing and shooting process is measured and analyzed according to the empirical analysis. The noise layer network mainly has: geometric distortion, motion blur, color transformation, noise attack and JEPG compression, wherein the geometric distortion and the motion blur are used for simulating attacks received during shooting, the intensity of the noise attack is a random value, and the value range is set to be changed according to environmental changes. Because the ceramic trademark is subjected to stronger color conversion attack during firing, a larger value range is set for the color conversion attack, because the ceramic watermark image can be attached to the ceramic only by high-temperature firing, and through the experience of actual firing, the pattern (watermark image) attached to the ceramic has certain color after being fired at high temperature generally, so that the strength of the color conversion subjected to high temperature during firing of the ceramic watermark image is stronger, and the range for simulating the color attack is synchronously enlarged to ensure that the secret information can be correctly extracted by a current decoder after the ceramic watermark image subjected to the color attack with the strength.
S106: sending the watermark image subjected to noise processing into a current decoder for decoding to obtain secret information, obtaining a loss value of a cross entropy loss function according to the secret information and the training watermark information, and updating the current decoder according to the loss value of the cross entropy loss function to obtain an updated current decoder.
Further, before updating the current decoder according to the loss value of the cross entropy loss function, the method further includes: and acquiring the weight value of the cross entropy loss function. Updating the current decoder according to the loss value of the cross-entropy loss function comprises: and updating the current decoder by using the weight value of the cross entropy loss function and the loss value of the cross entropy loss function.
In embodiment 1 of the present invention, the end condition of the training of the boundary region binarization ceramic watermark model may be a second step threshold, where the second step threshold may be determined according to training conditions of the encoder and the decoder, and specifically, the second step threshold is a training step number when the encoder reaches a preset first convergence condition and the decoder reaches a preset second convergence condition. Wherein the first convergence condition may be that the median of the binary pixel selection map is either 0 or 1; the second convergence condition is that the watermark image after passing through the noise layer can be correctly extracted by the decoder.
The overall network convergence method is as follows:
only the cross entropy loss function is reserved for training in the initial training stage, and the training step length is set to be about 2000 steps to 3000 steps. And when the average value of the loss values of the binary pixel selection graph loss function is about 0.9 to 1.0, performing breakpoint training (disconnecting the program and only adjusting parameters without changing the network structure), performing second-time gradient clipping of the binary pixel selection graph, setting the clipping range of the loss values of the binary pixel selection graph loss function to be (0, 1), and then optimizing the total loss function. The total step length is thirty thousand steps, and through the training scheme, the final model is converged, and the decoding rate is 95%.
Example 2
Corresponding to the embodiment 1 of the invention, the invention provides a ceramic watermark model training device with binaryzation of a boundary area. Fig. 11 is a schematic structural diagram of a ceramic watermark model training apparatus with boundary region binarization in embodiment 2 of the present invention. As shown in fig. 11, the training apparatus for a ceramic watermark model with binarized boundary region in embodiment 2 of the present invention includes an obtaining module 20, a preprocessing module 21, a watermark generating module 22, a first adjusting module 23, a noise processing module 24, and a second adjusting module 25.
Specifically, the obtaining module 20 is configured to obtain a training image and training watermark information respectively;
the preprocessing module 21 is configured to perform a corrosion operation on the training image to obtain a training corrosion image, and perform an expansion operation on the training image to obtain a training expansion image;
a watermark generating module 22, configured to input the training image and the training watermark information into the current encoder to generate a binary pixel selection map, and obtain a watermark image according to the binary pixel selection map, the training erosion image, and the training expansion image;
the first adjusting module 23 is configured to calculate a loss value of a binary pixel selection map loss function, and adjust the current encoder according to the loss value of the binary pixel selection map loss function to obtain an updated current encoder;
the noise processing module 24 is configured to place the watermark image in a preset noise layer for noise processing;
and the second adjusting module 25 is configured to send the watermark image subjected to noise processing to a current decoder for decoding to obtain secret information, obtain a cross-entropy loss function loss value according to the secret information and the training watermark information, and update the current decoder according to the cross-entropy loss function loss value to obtain an updated current decoder.
Specifically, the noise layer includes: geometric distortion, motion blur, color shift, gaussian noise, and JEPG compression.
More specifically, the distortion coefficient of the geometric distortion is less than 1; and/or, the motion blur adopts a linear blur kernel, the pixel width of the linear kernel is not more than 10, the linear angle is randomly selected, and the range is not more than 1/2 pi; and/or the offset value of the color offset is required to be uniformly distributed, and the offset value is-0.2-0.3; and/or, the compression quality factor of the JEPG compression is greater than 50.
The specific details of the training apparatus for the ceramic watermark model with binarized boundary region may be understood by referring to the corresponding relevant description and effects in the embodiments shown in fig. 1 to fig. 10, and are not described herein again.
Example 3
The embodiment 3 of the invention provides an encoder, which is obtained by training by using the ceramic watermark model training method with boundary area binarization described in the embodiment 1 of the invention.
Example 4
Embodiment 4 of the present invention provides a decoder, which is obtained by training using the boundary region binarization ceramic watermark model training method described in embodiment 1 of the present invention.
Example 5
The embodiment 5 of the invention provides an embedding method of a ceramic watermark model with binaryzation in a boundary area. The embedding and densifying method of the ceramic in the embodiment 5 of the invention comprises the following steps:
s501: and respectively acquiring training images and training watermark information.
S502: and carrying out corrosion operation on the training image to obtain a corrosion image, and carrying out expansion operation on the training image to obtain an expansion image.
S503: inputting the training image and the training watermark information into an encoder of embodiment 3 of the present invention to encode to obtain a binary pixel selection map, and obtaining an electronic watermark image by using the binary pixel selection map, the erosion selection map, and the expansion selection map;
s504: and after the electronic watermark image is transferred to the ceramic prefabricated product, firing the ceramic prefabricated product to obtain the ceramic with the watermark image.
As specific embodiments, the electronic watermark image can be transferred to the ceramic preform in two ways: inputting the electronic watermark image into a preset ceramic ink-jet injection machine, and carrying out ink-jet on the ceramic prefabricated product by using the ceramic ink-jet injection machine so as to transfer the electronic watermark image onto the ceramic prefabricated product; or generating paper edition stained paper according to the electronic watermark image; and paving the paper pattern paper on the ceramic prefabricated product so as to transfer the electronic watermark image onto the ceramic prefabricated product.
In a specific embodiment, when the ceramic preform is a commodity ceramic preform, firing the ceramic preform to obtain a ceramic with a watermark pattern comprises: firing the daily ceramic prefabricated product at 800-1380 ℃ to obtain daily ceramic; when the ceramic preform is a sanitary ceramic preform, firing the ceramic preform to obtain a ceramic having a watermark pattern comprises: firing the sanitary ceramic prefabricated product at 800-1380 ℃ to obtain sanitary ceramic; when the ceramic preform is an architectural ceramic preform, firing the ceramic preform to obtain a ceramic with a watermark pattern comprising: and firing the architectural ceramic prefabricated product at 800-1380 ℃ to obtain the architectural ceramic.
For example, fig. 12 is a flow chart of a method for manufacturing a ceramic watermark pattern based on an inkjet process, and as shown in fig. 12, a ceramic electronic trademark or pattern is first given, copyright watermark information is embedded into the electronic trademark (or pattern) by using a robust watermarking technology based on a digital image, so as to form a trademark containing the copyright information, then the trademark containing the copyright information is sent to a ceramic inkjet injector to color a ceramic carrier, and then the colored ceramic carrier is sent to a kiln to be fired at a high temperature, so as to finally form the ceramic carrier containing the copyright information. Fig. 13 is a flowchart of a method for manufacturing a ceramic watermark pattern based on screen printing, and as shown in fig. 13, an electronic version ceramic trademark or pattern is first given, and copyright information is embedded according to a robust watermarking technology to form an electronic version trademark pattern containing the copyright information. Then, generating paper plate stained paper (special paper for decorating ceramic) by relying on the electronic plate watermark picture, wherein the step of forming the paper plate stained paper comprises the following working procedures: making plate with stained paper, printing plate, mixing colors and making sample. Then the paper pattern paper containing copyright information is spread on the ceramic and is put into a kiln for firing. Finally, the patterns of the copyrighted stained paper fired by the kiln can be completely transferred to the ceramic, so that the copyright protection of the ceramic is realized.
Example 6
s601: the watermark pattern on the ceramic is located.
S602: and inputting the positioned watermark pattern into a decoder of embodiment 4 of the invention for decoding to obtain the watermark information in the watermark pattern.
As a specific implementation manner, the decryption method of the ceramic watermark pattern may adopt the following technical scheme: firstly, positioning and detecting the watermark pattern on the ceramic product by a high-precision scanner or a picture camera, then correcting the size of the positioned and detected picture, sending the corrected picture into a mobile phone or a computer, and then extracting the copyright information in the corrected picture by means of a robust watermark extraction algorithm in the mobile phone or the computer. And finally, comparing the copyright information content to judge whether the ceramic is infringed or not so as to achieve the function of copyright authentication.
For example, the copyright information content can be arbitrarily designed to form a watermark according to the intention of an author, such as the name of the author, company information, brand name, ceramic number and the like, so as to prove that the ceramic copyright belongs to. And then embedding the watermark into a ceramic trademark or a pattern prepared in advance by using a robust watermark algorithm to obtain an electronic version watermark picture containing the watermark. Fig. 14 is a schematic flow chart of the ceramic copyright encryption and decryption based on the ink-jet process, if the ink-jet process is adopted, the electronic version watermark picture is directly sent to a ceramic ink-jet machine to print and color a ceramic carrier, and then the ceramic carrier is sent to a kiln to be fired at 1170 ℃ to obtain the ceramic product containing the copyright information. Fig. 15 is a schematic flow chart of ceramic copyright encryption and decryption based on screen printing, in case of a screen printing process, an electronic version watermark picture is formed by the steps of pattern making of stained paper, plate burning, color mixing, sample preparation and the like, then the ceramic processes of over-glaze, in-glaze and under-glaze are selected according to different application scenes of the ceramic product, after the corresponding ceramic process is selected, the manufactured paper version watermark picture and a ceramic carrier are put into a kiln to be fired, and finally the ceramic product containing copyright information is obtained.
The method comprises the following steps of screening copyright information after a customer purchases a ceramic product:
firstly, positioning and detecting trademarks or patterns on the ceramic product through a high-precision scanner or a picture camera, correcting the size of the detected picture, then putting the corrected picture into a mobile phone or a computer transplanted with a robust watermark extraction algorithm to extract copyright information, and then comparing the content of the copyright information to judge whether the ceramic product is infringed so as to achieve the function of copyright authentication.
Example 7
Embodiments of the present invention further provide an electronic device, which may include a processor and a memory, where the processor and the memory may be connected by a bus or in another manner.
The processor may be a Central Processing Unit (CPU). The Processor may also be other general purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field Programmable Gate Arrays (FPGAs) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, or a combination thereof.
The memory is a non-transitory computer readable storage medium, and may be used to store a non-transitory software program, a non-transitory computer executable program, and modules, such as program instructions/modules (for example, the obtaining module 20, the preprocessing module 21, the watermark generating module 22, the first adjusting module 23, the noise processing module 24, and the second adjusting module 25 shown in fig. 11) corresponding to the ceramic watermark model training method for number boundary area binarization in the embodiment of the present invention.
The memory may include a storage program area and a storage data area, wherein the storage program area may store an operating system, an application program required for at least one function; the storage data area may store data created by the processor, and the like. Further, the memory may include high speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid state storage device. In some embodiments, the memory optionally includes memory located remotely from the processor, and such remote memory may be coupled to the processor via a network. Examples of such networks include, but are not limited to, the internet, intranets, local area networks, mobile communication networks, and combinations thereof.
The one or more modules are stored in the memory, and when executed by the processor, perform a ceramic watermark model training method with boundary region binarization as in the embodiments shown in fig. 1 to 10.
The details of the electronic device may be understood by referring to the corresponding descriptions and effects in the embodiments shown in fig. 1 to fig. 11, and are not described herein again.
Those skilled in the art will appreciate that all or part of the processes in the methods of the embodiments described above can be implemented by instructing relevant hardware by a computer program, where the computer program can be stored in a computer-readable storage medium, and when executed, the computer program can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic Disk, an optical Disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), a Flash Memory (Flash Memory), a Hard Disk (Hard Disk Drive, abbreviated as HDD) or a Solid State Drive (SSD), etc.; the storage medium may also comprise a combination of memories of the kind described above.
Although the embodiments of the present invention have been described in conjunction with the accompanying drawings, those skilled in the art may make various modifications and variations without departing from the spirit and scope of the invention, and such modifications and variations fall within the scope defined by the appended claims.
Claims (14)
1. A ceramic watermark model training method of boundary region binaryzation is characterized by comprising the following steps:
respectively acquiring training images and training watermark information;
carrying out corrosion operation on the training image to obtain a training corrosion image, and carrying out expansion operation on the training image to obtain a training expansion image;
inputting the training image and the training watermark information into a current encoder to generate a binary pixel selection graph, and obtaining a watermark image according to the binary pixel selection graph, the training erosion image and the training expansion image;
calculating the current loss value of the binary pixel selection graph loss function, and adjusting the current encoder according to the current loss value of the binary pixel selection graph loss function to obtain an updated current encoder;
putting the watermark image into a preset noise layer for noise processing;
sending the watermark image subjected to noise processing into a current decoder for decoding to obtain secret information, obtaining a loss value of a cross entropy loss function according to the secret information and the training watermark information, and updating the current decoder according to the loss value of the cross entropy loss function to obtain an updated current decoder;
obtaining a watermark image according to the binary pixel selection map, the training erosion image and the training expansion image comprises:
performing dot multiplication on the binary pixel selection image and the training corrosion image to obtain a corrosion selection image;
performing reverse color processing on the binary pixel selection image, and performing dot multiplication on the binary pixel selection image subjected to the reverse color processing and the training expansion image to obtain an expansion selection image;
superposing the corrosion selection diagram and the expansion selection diagram to obtain the watermark image;
the binary pixel selection image is a binary pixel selection image, and is a single-channel binary image with the same size as the training erosion image and the training expansion image, and the binary pixel selection image has the function of selecting corresponding pixels in the training erosion image or the training expansion image to output according to needs.
2. The method of claim 1, further comprising, prior to adjusting the current encoder based on the loss value of the binary pixel selection map loss function: obtaining a weight value of the loss function of the binary pixel selection graph;
adjusting the current encoder according to the loss value of the binary pixel selection map loss function comprises: updating the current encoder by using the weight value of the binary pixel selection graph loss function and the loss value of the binary pixel selection graph loss function;
before updating the current decoder according to the loss value of the cross-entropy loss function, the method further comprises: acquiring a weight value of the cross entropy loss function;
updating the current decoder according to the loss value of the cross entropy loss function comprises: and updating the current decoder by using the weight value of the cross entropy loss function and the loss value of the cross entropy loss function.
3. The method of claim 2,
before a preset first step threshold value, the weight value of the binary pixel selection graph loss function is 0; after the first step number threshold, a weight value of the binary pixel selection map loss function is greater than 0 and less than a weight value of the cross entropy loss function.
4. The method of claim 3, further comprising, prior to obtaining the training image and the training watermark information separately: and carrying out first numerical clipping on the loss function value of the binary pixel selection map.
5. The method of claim 4, further comprising, after said computing the current penalty value for the binary pixel selection map penalty function:
judging whether the pixel distribution in the binary pixel selection image reaches a preset condition or not according to the current loss value of the binary pixel selection image loss function;
and performing second numerical clipping on the loss function value of the binary pixel selection image after the pixel distribution in the binary pixel selection image reaches a preset condition.
6. A ceramic watermark model training device with binaryzation in a boundary area is characterized by comprising:
the acquisition module is used for respectively acquiring training images and training watermark information;
the preprocessing module is used for carrying out corrosion operation on the training image to obtain a training corrosion image and carrying out expansion operation on the training image to obtain a training expansion image;
the watermark generating module is used for inputting the training image and the training watermark information into a current encoder to generate a binary pixel selection image, and obtaining a watermark image according to the binary pixel selection image, the training erosion image and the training expansion image; obtaining a watermark image according to the binary pixel selection map, the training erosion image and the training expansion image comprises: performing dot multiplication on the binary pixel selection image and the training corrosion image to obtain a corrosion selection image; performing reverse color processing on the binary pixel selection image, and performing dot multiplication on the binary pixel selection image subjected to the reverse color processing and the training expansion image to obtain an expansion selection image; superposing the corrosion selection diagram and the expansion selection diagram to obtain the watermark image;
the first adjusting module is used for calculating a loss value of a binary pixel selection map loss function, and adjusting the current encoder according to the loss value of the binary pixel selection map loss function to obtain an updated current encoder;
the noise processing module is used for putting the watermark image into a preset noise layer for noise processing;
the second adjusting module is used for sending the watermark image subjected to noise processing into a current decoder for decoding to obtain secret information, obtaining a loss value of a cross entropy loss function according to the secret information and the training watermark information, and updating the current decoder according to the loss value of the cross entropy loss function to obtain an updated current decoder;
the binary pixel selection image is a binary pixel selection image, and is a single-channel binary image with the same size as the training erosion image and the training expansion image, and the binary pixel selection image has the function of selecting corresponding pixels in the training erosion image or the training expansion image to output according to needs.
7. The apparatus of claim 6, wherein the noise layer comprises: geometric distortion, motion blur, color shift, gaussian noise, and JEPG compression.
8. The apparatus of claim 7, wherein:
the distortion coefficient of the geometric distortion is less than 1;
and/or the motion blur adopts a linear blur kernel, the pixel width of the linear blur kernel is not more than 10, the linear angle is randomly selected, and the range is not more than 1/2 pi;
and/or the offset value of the color offset is required to be uniformly distributed, and the offset value is-0.2-0.3;
and/or, the compression quality factor of the JEPG compression is greater than 50.
9. An encoder, characterized in that, the encoder is obtained by training with the ceramic watermark model training method of boundary area binarization as claimed in any one of claims 1-5.
10. A decoder is characterized by being obtained by training through the ceramic watermark model training method with the binaryzation of the boundary area according to any one of claims 1 to 5.
11. A ceramic watermark embedding method of boundary region binaryzation is characterized by comprising the following steps:
respectively acquiring training images and training watermark information;
carrying out corrosion operation on the training image to obtain a corrosion image, and carrying out expansion operation on the training image to obtain an expansion image;
inputting the training image and the training watermark information into the encoder of claim 9 for encoding to obtain a binary pixel selection map, and obtaining an electronic watermark image by using the binary pixel selection map, the erosion selection map and the expansion selection map;
and after the electronic watermark image is transferred to the ceramic prefabricated product, firing the ceramic prefabricated product to obtain the ceramic with the watermark pattern.
12. The method according to claim 11, wherein transferring the electronic watermark image onto the ceramic preform comprises:
inputting the electronic watermark image into a preset ceramic ink-jet injection machine, and carrying out ink-jet on the ceramic prefabricated product by using the ceramic ink-jet injection machine so as to transfer the electronic watermark image onto the ceramic prefabricated product;
or, generating paper edition stained paper according to the electronic watermark image;
and (3) paving the paper pattern paper on the ceramic prefabricated product to transfer the electronic watermark image to the ceramic prefabricated product.
13. The method of claim 11, wherein:
when the ceramic preform is a commodity ceramic preform, firing the ceramic preform to obtain a ceramic having a watermark pattern comprising:
firing the daily ceramic prefabricated product at 800-1380 ℃ to obtain daily ceramic;
when the ceramic preform is a sanitary ceramic preform, firing the ceramic preform to obtain a ceramic having a watermark pattern comprises:
firing the sanitary ceramic prefabricated product at 800-1380 ℃ to obtain sanitary ceramic;
when the ceramic preform is an architectural ceramic preform, firing the ceramic preform to obtain a ceramic with a watermark pattern comprising:
and firing the architectural ceramic prefabricated product at 800-1380 ℃ to obtain the architectural ceramic.
14. A method for decrypting a ceramic watermark, comprising:
positioning the watermark pattern on the ceramic;
inputting the positioned watermark pattern into the decoder of claim 10 for decoding, and obtaining the training watermark information in the watermark pattern.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110844701.3A CN113837915B (en) | 2021-07-26 | 2021-07-26 | Ceramic watermark model training method and embedding method for binaryzation of boundary region |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110844701.3A CN113837915B (en) | 2021-07-26 | 2021-07-26 | Ceramic watermark model training method and embedding method for binaryzation of boundary region |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN113837915A CN113837915A (en) | 2021-12-24 |
| CN113837915B true CN113837915B (en) | 2022-09-23 |
Family
ID=78962949
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110844701.3A Active CN113837915B (en) | 2021-07-26 | 2021-07-26 | Ceramic watermark model training method and embedding method for binaryzation of boundary region |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN113837915B (en) |
Family Cites Families (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111491170B (en) * | 2019-01-26 | 2021-12-10 | 华为技术有限公司 | Method for embedding watermark and watermark embedding device |
| KR102412135B1 (en) * | 2019-08-01 | 2022-06-23 | 네이버 주식회사 | Image synthesizing apparatus and method for inserting watermark |
| CN113052745B (en) * | 2021-04-25 | 2022-01-07 | 景德镇陶瓷大学 | Digital watermark model training method, ceramic watermark image manufacturing method and ceramic |
-
2021
- 2021-07-26 CN CN202110844701.3A patent/CN113837915B/en active Active
Also Published As
| Publication number | Publication date |
|---|---|
| CN113837915A (en) | 2021-12-24 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US7599099B2 (en) | Image processing apparatus and image processing method | |
| CN113052745B (en) | Digital watermark model training method, ceramic watermark image manufacturing method and ceramic | |
| JP4218920B2 (en) | Image processing apparatus, image processing method, and storage medium | |
| US6993150B2 (en) | Halftone primitive watermarking and related applications | |
| US20010030759A1 (en) | Image processing apparatus for determining specific images | |
| CN113222804B (en) | Ceramic process-oriented up-sampling ceramic watermark model training method and embedding method | |
| CN101122995A (en) | Binary image digit water mark embedding, extraction method and device | |
| Villán et al. | Text data-hiding for digital and printed documents: Theoretical and practical considerations | |
| Mizumoto et al. | Robustness investigation of DCT digital watermark for printing and scanning | |
| CN105159627A (en) | Method for automatically correcting digital printing chromatic aberration | |
| Son et al. | Watermark detection from clustered halftone dots via learned dictionary | |
| CN113379585B (en) | Ceramic watermark model training method and embedding method for frameless positioning | |
| CN113538201B (en) | Ceramic watermark model training method and device based on bottom changing mechanism and embedding method | |
| CN112217958B (en) | Method for preprocessing digital watermark carrier image irrelevant to device color space | |
| CN113837915B (en) | Ceramic watermark model training method and embedding method for binaryzation of boundary region | |
| CN106327416B (en) | A kind of site water mark method based on printed matter | |
| KR102158633B1 (en) | A method for extraction of a registered seal impression from a document | |
| Son et al. | Inverse color to black-and-white halftone conversion via dictionary learning and color mapping | |
| JP4311698B2 (en) | Image processing apparatus, image processing method, and recording medium | |
| JP3809310B2 (en) | Image processing apparatus and method, and storage medium | |
| WO2020119650A1 (en) | Reversible watermarking method | |
| Fu et al. | Data hiding in halftone images by conjugate error diffusion | |
| JP4262219B2 (en) | Image processing apparatus and control method thereof | |
| JP3869983B2 (en) | Image processing apparatus and method, and storage medium | |
| CN110570343A (en) | Image watermark embedding method and device based on self-adaptive feature point extraction |
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 |