WO2021083110A1

WO2021083110A1 - Carrier object processing and watermark embedding methods and apparatuses, and electronic device

Info

Publication number: WO2021083110A1
Application number: PCT/CN2020/123884
Authority: WO
Inventors: 崔文学; 张琦; 刘永亮
Original assignee: 阿里巴巴集团控股有限公司
Priority date: 2019-10-31
Filing date: 2020-10-27
Publication date: 2021-05-06
Also published as: CN112750065A; CN112750065B

Abstract

Carrier object processing and watermark embedding methods and apparatuses, and an electronic device. The carrier object processing method comprises: acquiring a carrier object to be processed (S201); extracting a watermark template from said carrier object (S202); editing a watermark in said carrier object according to the watermarking template, so as to obtain a target carrier object (S203). According to the described method, the specific position of the watermark can be accurately obtained by extracting the watermark template corresponding to a carrier object to be processed, and the watermark of said carrier object can be accurately edited according to the watermark template, solving the problem in the existing watermark editing method that the watermark cannot be accurately edited due to diversity of watermark patterns.

Description

Carrier object processing and watermark embedding method, device and electronic equipment

This application claims the priority of the Chinese patent application filed on October 31, 2019 with the application number 201911049751.1 and the invention title "Carrier object processing and watermark embedding method, device and electronic equipment", the entire content of which is incorporated into this application by reference in.

Technical field

This application relates to the field of computer technology, in particular to a method, device, and electronic device for processing carrier objects, and a method, device, and electronic device for embedding a watermark.

Background technique

In recent years, with the rapid development of information technology and digital technology, more and more digital media, including sound, text, images, video, etc., have been widely disseminated. These digital media have the characteristics of easy storage, copying and dissemination, and are easy to carry out various forms of operation and modification, thus making these digital media easy to be used by criminals in the process of dissemination. Therefore, more and more companies and individuals are adding watermarks to these digital media to protect their intellectual property rights.

In some application scenarios, it is necessary to edit the watermark carried by the digital media, which involves watermark editing technology. The existing watermark editing technology is mainly to obtain a large number of carrier objects that contain the same or similar watermark, and then estimate the watermark position of the carrier object, and then edit the watermark according to the estimated watermark position. However, using this technology can only process a batch of carrier objects, and cannot process a single carrier object alone, which brings troubles to the processing of a single carrier object. In addition, the above-mentioned watermark editing technique can only be edited for a certain type of watermark, that is, only a single type of watermark can be edited. However, the types of watermarks in the existing watermarked carrier objects are different. Therefore, the existing watermark editing methods cannot accurately edit the watermarks in the carrier objects with watermarks with diversified styles.

Summary of the invention

This application provides a carrier object processing method to solve the problem that the existing watermark editing methods cannot accurately edit the watermark due to the diversity of watermark styles; this application also provides a carrier object processing device and electronic equipment. This application also provides a watermark embedding method, device and electronic equipment.

This application provides a method for processing a carrier object, including:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

According to the watermark template, edit the watermark in the to-be-processed carrier object to obtain a target carrier object.

Optionally, the extracting a watermark template from the to-be-processed carrier object includes:

Input the to-be-processed carrier object into a network model for extracting a watermark template to obtain a feature vector of the to-be-processed carrier object;

According to the feature vector of the carrier object to be processed, a watermark template corresponding to the feature vector is obtained.

Optionally, the obtaining the watermark template corresponding to the feature vector according to the feature vector of the carrier object to be processed includes:

Performing a convolution operation on the feature vector to obtain a watermark template corresponding to the feature vector.

Optionally, the watermark template is a template used when adding watermark information to the original carrier object; wherein, the original carrier object is the carrier object before the watermark information is added to the carrier object to be processed.

Optionally, it also includes:

Setting the color of the watermark in the carrier object to be processed to a predetermined color according to the watermark template;

The editing the watermark in the carrier object to be processed according to the watermark template to obtain a target carrier object includes: editing in the carrier object to be processed according to a predetermined color of the watermark and the watermark template The watermark obtains the target carrier object.

Optionally, the setting the color of the watermark to a predetermined color in the carrier object to be processed according to the watermark template includes:

The carrier object to be processed and the watermark template are input into a network model for setting the watermark color to a predetermined color, so as to set the color of the watermark in the carrier object to be processed to a predetermined color.

Optionally, the input of the carrier object to be processed and the watermark template into a network model for setting the watermark color to a predetermined color, so as to set the color of the watermark in the carrier object to be processed to a predetermined color ,include:

Obtaining the position information and shape information of the watermark in the carrier object to be processed according to the watermark template;

According to the position information and shape information of the watermark in the carrier object to be processed, the color of the watermark in the carrier object to be processed is set to a predetermined color.

Optionally, it also includes:

Obtaining the carrier object in which the color of the watermark in the carrier object to be processed is set to a predetermined color;

The editing the watermark in the to-be-processed carrier object according to the predetermined color of the watermark and the watermark template to obtain a target carrier object includes:

The carrier object to be processed, the watermark template, and the carrier object set to a predetermined color are input into the network model for editing the watermark to obtain the target carrier object.

Optionally, the inputting the to-be-processed carrier object, the watermark template, and the carrier object set to a predetermined color into a network model for editing watermarks to obtain a target carrier object includes:

Input the to-be-processed carrier object, the watermark template, and the carrier object set to a predetermined color into the network model for editing watermarks to obtain the to-be-processed carrier object, the watermark template, and the The edit feature vector corresponding to the carrier object set as a predetermined color;

According to the to-be-processed carrier object, the watermark template, and the editing feature vector corresponding to the carrier object set to a predetermined color, edit the watermark of the to-be-processed carrier object to obtain a target carrier object.

Optionally, the setting the color of the watermark in the to-be-processed carrier object to a predetermined color refers to setting the pixel value of the watermark in the to-be-processed carrier object at the position in the to-be-processed carrier object Is the predetermined pixel value.

Optionally, it further includes obtaining the network model used to extract the watermark template;

The obtaining of the network model used to extract the watermark template includes:

Obtain sample data for extracting the watermark template;

Obtaining the first-level feature vector of the sample data;

Performing a convolution operation on the first-level feature vector to obtain a first-level watermark template corresponding to the first-level feature vector;

Obtaining a second-level feature vector according to the first-level feature vector;

Performing a convolution operation on the secondary feature vector to obtain a secondary watermark template corresponding to the secondary feature vector;

Obtaining a third-level feature vector according to the second-level feature vector;

Performing a convolution operation on the three-level feature vector to obtain a three-level watermark template corresponding to the three-level feature vector;

Using the first-level watermark template, the second-level watermark template, and the third-level watermark template as watermark templates corresponding to the sample data;

The corresponding relationship between the sample data and the watermark template corresponding to the sample data is obtained, and the corresponding relationship is used as a network model for extracting the watermark template.

Optionally, the obtaining the second-level feature vector according to the first-level feature vector includes:

Obtaining, according to the first-level feature vector, the feature vector of the first shrinkage scale corresponding to the first-level feature vector;

Obtaining the feature vector of the first contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector;

The second-level feature vector is obtained according to the feature vector of the first shrinkage scale corresponding to the second-level feature vector.

Optionally, the obtaining the feature vector of the first contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector includes:

Obtaining the feature vector of the second contraction scale corresponding to the first-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector;

Obtaining the feature vector of the second contraction scale corresponding to the second-level feature vector according to the feature vector of the second contraction scale corresponding to the first-level feature vector;

According to the feature vector of the second contraction scale corresponding to the second-level feature vector, the feature vector of the first contraction scale corresponding to the second-level feature vector is obtained.

Optionally, the obtaining a third-level feature vector according to the second-level feature vector includes:

Obtaining, according to the second-level feature vector, a feature vector of the first shrinkage scale corresponding to the second-level feature vector;

Obtaining the feature vector of the first contraction scale corresponding to the third-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector;

According to the feature vector of the first shrinkage scale corresponding to the three-level feature vector, a three-level feature vector is obtained.

Optionally, the obtaining the feature vector of the first contraction scale corresponding to the third-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector includes:

Obtaining the feature vector of the second contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector;

Obtaining the feature vector of the second contraction scale corresponding to the third-level feature vector according to the feature vector of the second contraction scale corresponding to the second-level feature vector;

According to the feature vector of the second contraction scale corresponding to the three-level feature vector, the feature vector of the first contraction scale corresponding to the three-level feature vector is obtained.

Optionally, it also includes obtaining a network model for setting the watermark color to a predetermined color;

The obtaining a network model for setting the watermark color to a predetermined color includes:

According to the first-level feature vector and the first-level watermark template, set the color of the watermark in the sample data to a predetermined color, and obtain a first sample of the sample data set to the predetermined color;

Setting the color of the watermark in the sample data to a predetermined color according to the secondary feature vector and the secondary watermark template; obtaining a second sample of the sample data set to the predetermined color;

According to the three-level feature vector and the three-level watermark template, set the color of the watermark in the sample data to a predetermined color; obtain a third sample of the sample data set to the predetermined color;

Using the first sample, the second sample, and the third sample as sample data in which the color of the watermark is set to a predetermined color in the sample data;

The corresponding relationship between the sample data and the sample data of the predetermined color is obtained, and the corresponding relationship is used as a network model for setting the watermark color to the predetermined color.

Optionally, the method further includes: correcting the network model for extracting the watermark template according to the sample data of the predetermined color;

The correcting the network model for extracting the watermark template according to the sample data of the predetermined color includes:

Determine whether the sample data of the predetermined color and the sample data of the desired predetermined color are within the specified threshold; if so, directly use the sample data of the predetermined color as the network model for setting the watermark color to the predetermined color Output result;

If not, reacquire the correction feature vector of the carrier object to be processed according to the sample data of the predetermined color and the sample data used to extract the watermark template; obtain the correction feature vector according to the correction feature vector Corresponding corrected watermark template.

Optionally, it also includes obtaining a network model for editing the watermark;

The obtaining of the network model used for editing the watermark includes:

Obtaining the edit feature vector in the sample data according to the sample data, the watermark template corresponding to the sample data, and the sample data of the predetermined color;

According to the editing feature vector, edit the watermark corresponding to the sample data to obtain target sample data;

The corresponding relationship between the sample data and the target sample data is obtained, and the corresponding relationship is used as a network model for editing the watermark.

Automatically obtain at least one set of watermark templates corresponding to the to-be-processed carrier object according to the to-be-processed carrier object;

A set of watermark templates or all of the watermark templates in the at least one set of watermark templates corresponding to the carrier object to be processed are used as the watermark templates of the carrier to be processed.

Correspondingly, this application provides a carrier object processing device, including:

The carrier object processing unit is used to obtain the carrier object to be processed;

A watermark template extraction unit, configured to extract a watermark template from the carrier object to be processed;

The target carrier object obtaining unit is configured to edit the watermark in the to-be-processed carrier object according to the watermark template to obtain the target carrier object.

This application also provides a watermark embedding method, including:

Obtain the original carrier object and obtain the original watermark;

Obtaining position information at which the original watermark is embedded on the original carrier object;

In the original carrier object, randomly according to the attribute information of the original watermark and the position information, the original watermark is embedded in a designated position of the original carrier object.

Optionally, the obtaining the original watermark includes:

Randomly combine at least one of numeric symbols, English symbols, and special symbols to form a character sequence;

Use the character sequence as the original watermark.

Optionally, the obtaining the position information of the original watermark embedded on the original carrier object includes:

Acquiring the size information of the original carrier object and the size information of the original watermark;

Determine the embedding area of the original watermark on the original carrier object according to the size information of the original carrier object;

In the embedding area, according to the size information of the original watermark, the position information at which the original watermark is embedded on the original carrier object is obtained.

Optionally, the determining the embedding area of the original watermark on the original carrier object according to the size information of the original carrier object includes:

Obtaining the length and size information of the original carrier object according to the size information of the original carrier object;

Determine the first reserved size information in the length direction on the original carrier object according to the length size information;

Obtaining the width size information of the original carrier object according to the size information of the original carrier object;

Determine the second reserved size information in the width direction on the original carrier object according to the width size information;

According to the size information of the original carrier object, the first reserved size information and the second reserved size information, the embedding area of the original watermark on the original carrier object is determined.

Optionally, the obtaining, in the embedding area, the position information at which the original watermark is embedded on the original carrier object according to the size information of the original watermark, includes:

Obtaining edge position information of the original watermark according to the size information of the original watermark;

According to the edge position information of the original watermark and the position information of the embedding area, the position information at which the original watermark is embedded on the original carrier object is obtained.

Optionally, the attribute information of the original watermark includes at least one of the color of the original watermark, the font of the original watermark, and the font size information of the original watermark.

Correspondingly, this application also provides a watermark embedding device, including:

The original carrier object obtaining unit, which is used to obtain the original carrier object;

The original watermark obtaining unit is used to obtain the original watermark;

A location information obtaining unit, configured to obtain location information of the original watermark embedded on the original carrier object;

The embedding unit is configured to embed the original watermark at a designated position of the original carrier object randomly according to the attribute information of the original watermark and the position information in the original carrier object.

This application provides an electronic device, including:

processor;

The memory is used to store the program of the carrier object processing method and execute the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

This application provides an electronic device, including:

processor;

The memory is used to store the program of the watermark embedding method and execute the following steps:

Obtain the original carrier object and obtain the original watermark;

This application provides a computer storage medium that stores a program of a carrier object processing method, and the program is run by a processor to perform the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

This application provides a computer storage medium that stores a program of the watermark embedding method, and the program is run by a processor to perform the following steps:

Obtain the original carrier object and obtain the original watermark;

Compared with the prior art, this application has the following advantages:

This application discloses a method for processing a carrier object, including: obtaining a carrier object to be processed; extracting a watermark template from the carrier object to be processed; editing the watermark in the carrier object to be processed according to the watermark template, Get the target carrier object. In the carrier object processing method of the present application, by extracting the watermark template corresponding to the carrier object to be processed, the specific position of the watermark can be accurately obtained; according to the watermark template, the watermark of the carrier object to be processed can be accurately edited. The carrier object processing method of the present application solves the problem that the existing watermark editing method cannot accurately edit the watermark due to the diversity of watermark styles by obtaining the watermark template corresponding to the carrier object to be processed.

Description of the drawings

In order to more clearly describe the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only These are some embodiments described in this application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings.

FIG. 1A is a schematic diagram of an application scenario embodiment of the carrier object processing method provided by this application.

FIG. 1B is a schematic diagram of another embodiment of the application scenario of the carrier object processing method provided by this application.

FIG. 2 is a flowchart of a method for processing a carrier object according to Embodiment 1 of the application.

FIG. 3 is a schematic diagram of a second application scenario embodiment of the carrier object processing method provided by this application.

Fig. 4 is a schematic diagram of the GridNet network structure provided in the first embodiment of the application.

FIG. 5 is a schematic diagram of a carrier object processing apparatus provided in Embodiment 2 of this application.

Fig. 6 is a flow chart of the watermark embedding method provided in the third embodiment of the application.

Fig. 7 is a schematic diagram of a watermark embedding device provided in the fourth embodiment of the application.

FIG. 8 is a schematic diagram of a carrier object processing electronic device provided in Embodiment 5 of this application.

Detailed ways

In the following description, many specific details are explained in order to fully understand the present invention. However, the present invention can be implemented in many other ways different from those described here, and those skilled in the art can make similar popularizations without violating the connotation of the present invention. Therefore, the present invention is not limited by the specific implementation disclosed below.

This application provides a method and device for processing a carrier object, an electronic device, and a computer storage medium.

Some embodiments provided in this application can handle the scene of the watermark in the carrier object. As shown in FIG. 1A, it is a schematic diagram of an application scenario embodiment of the carrier object processing method provided by this application. First, the server obtains the carrier object to be processed from the terminal. The carrier object to be processed may be an image with a watermark or a video frame image. Extract the watermark template based on the current image. After obtaining the watermark template, set the color of the watermark in the watermarked image to a predetermined color, and then edit the watermark in the watermarked image according to the predetermined color of the watermark and the watermark template, and edit the watermarked image Return to the terminal. The editing of the watermark includes at least one of detecting the watermark, deleting the watermark, adding the watermark, and modifying the watermark. For example, FIG. 1B is a schematic diagram of another embodiment of the application scenario of the carrier object processing method provided by this application. In FIG. Add a new watermark "@blue#" to the images without watermark. It should be noted that the application scenario is only an embodiment of the application scenario, and the purpose of providing this application scenario embodiment is to facilitate the understanding of the carrier object processing method of the present application, and is not used to limit the carrier object processing method of the present application.

As shown in FIG. 2, it is a flowchart of an embodiment of a method for processing a carrier object according to Embodiment 1 of the present application. The method includes the following steps.

Step S201: Obtain a carrier object to be processed.

In this application, the carrier object to be processed refers to a carrier object containing a watermark. Wherein, the carrier object may be at least one of an image or a text file. For example, the carrier object containing the watermark can be an image containing a watermark, a Word text document containing a watermark, or a PDF text document containing a watermark. In this embodiment, the method of processing a carrier object is mainly described by processing a watermarked image. Of course, it is understandable that the carrier object to be processed may also be other carrier media containing watermarks other than images containing watermarks and text documents containing watermarks, and there is no limitation here.

Step S202: Extract a watermark template from the carrier object to be processed.

After obtaining the carrier object to be processed in step S201, in order to detect the precise position of the watermark in the carrier object to be processed, a watermark template needs to be extracted from the carrier object to be processed. The watermark template is a template used when adding watermark information to the original carrier object; wherein, the original carrier object is the carrier object before the watermark information is added to the carrier object to be processed.

Specifically, extracting the watermark template from the to-be-processed carrier object may be in the manner described below.

First, input the to-be-processed carrier object into the network model for extracting the watermark template, and obtain the feature vector of the to-be-processed carrier object.

The network model used to extract the watermark template is mainly used for the detection of the watermark in the carrier object to be processed, and the main task of the watermark detection is to extract the watermark binary template information from the watermarked carrier object (ie image), that is, the watermark template. In order to extract the watermark template more accurately, this embodiment designs a network model for extracting the watermark template. In the proposed network model for extracting the watermark template, a multi-cascade network structure is adopted. In the proposed multi-cascade network, each stage will output a feature vector F _i , where i represents the number of stages of the network, and then this feature vector will be used as the input feature vector of the next level (i+1) network. In this embodiment, the number of stages of the network is three. Each level of network is used for feature extraction (first level network), feature enhancement (second level network), and feature refinement (third level network).

After the feature vector is obtained, the watermark template corresponding to the feature vector is obtained according to the feature vector of the carrier object to be processed.

Specifically, the obtaining the watermark template corresponding to the feature vector according to the feature vector of the carrier object to be processed may be performing a convolution operation on the feature vector to obtain the watermark template corresponding to the feature vector. S _i output a watermark template convolution operation (C _i) on each stage output feature vector F _i. Since in the present embodiment, the three stages of the network, the output of the watermark templates S _i is also three.

In order to more accurately obtain the watermark template extracted from the carrier object to be processed, each level of the network in the network model for extracting the watermark template in this embodiment is a GridNet structure, as shown in Figure 4 below, which is a schematic diagram of the GridNet network structure. In the three network models in this embodiment, each network model uses a 3-cascade network, that is: L=3, the three-cascade network completes feature extraction, feature enhancement, and feature refinement in sequence, and each network The structure adopts the GridNet structure. Specifically, in GridNet, the convolution kernels of all convolution layers are 3x3, and the number of convolution kernels in the middle layer is 32, that is, 32 feature maps are generated for each layer. Down-sampling (down arrow in Figure 4) is implemented using convolution, with a step size of 2; up-sampling (up arrow in Figure 4) uses a deconvolution operation. In the structure in Figure 4, the horizontal arrows indicate that there are 3 convolutional layers, 3 activation layers, and 3 batch normalization layers; down-sampling arrows indicate 3 convolutional layers, 3 activation layers, and the last volume The build-up step size is 2; the up-sampling arrows indicate 2 convolutional layers, 1 deconvolutional layer, and 3 activation layers.

The GridNet network structure mainly uses the multi-scale information of the image, and the multi-scale information of the image can enhance the accuracy of extracting the watermark template from the carrier object to be processed to a certain extent, and can also enhance the learning ability of the network to a certain extent. If the i-th GridNet is expressed as

among them

Means

For the learnable parameters in, the entire network model used to extract the watermark template can be expressed as:

S _i =C _i (F _i ).

Before using the network model for extracting the watermark template to extract the template for the carrier object to be processed, it is necessary to obtain the network model for extracting the watermark template.

As one of the ways to obtain the network model for extracting the watermark template, the sample data for extracting the watermark template may be used first. Afterwards, the first-level feature vector of the sample data is obtained.

After obtaining the first-level feature vector of the sample data, perform a convolution operation on the first-level feature vector to obtain a first-level watermark template corresponding to the first-level feature vector.

At the same time, after the first-level feature vector of the sample data is obtained, the second-level feature vector is obtained according to the first-level feature vector. In a manner similar to obtaining the first-level watermark template corresponding to the first-level feature vector, a convolution operation is performed on the second-level feature vector to obtain the second-level watermark template corresponding to the second-level feature vector.

Similarly, after obtaining the second-level feature vector of the sample data, a third-level feature vector is obtained according to the second-level feature vector. In a manner similar to obtaining the second-level watermark template corresponding to the second-level feature vector, a convolution operation is performed on the third-level feature vector to obtain the third-level watermark template corresponding to the third-level feature vector.

After obtaining the first-level watermark template, the second-level watermark template, and the third-level watermark template, the first-level watermark template, the second-level watermark template, and the third-level watermark template are used as the watermark templates corresponding to the sample data.

Finally, the corresponding relationship between the sample data and the watermark template corresponding to the sample data is obtained, and the corresponding relationship is used as a network model for extracting the watermark template.

The GridNet structure adopted by each level of the network in the above-mentioned network model for extracting the watermark template may be specifically embodied in the manner of obtaining the second-level feature vector according to the first-level feature vector. Specifically, according to the first-level feature vector, the manner of obtaining the second-level feature vector is as described below.

First, according to the first-level feature vector, the feature vector of the first shrinkage scale corresponding to the first-level feature vector is obtained. Then, according to the feature vector of the first contraction scale corresponding to the first-level feature vector, the feature vector of the first contraction scale corresponding to the second-level feature vector is obtained. Finally, the second-level feature vector is obtained according to the feature vector of the first shrinkage scale corresponding to the second-level feature vector.

More specifically, according to the feature vector of the first contraction scale corresponding to the first-level feature vector, the feature vector of the first contraction scale corresponding to the second-level feature vector is obtained in the manner described below.

First, according to the feature vector of the first contraction scale corresponding to the first-level feature vector, the feature vector of the second contraction scale corresponding to the first-level feature vector is obtained. Then, according to the feature vector of the second contraction scale corresponding to the first-level feature vector, the feature vector of the second contraction scale corresponding to the second-level feature vector is obtained. Finally, according to the feature vector of the second contraction scale corresponding to the second-level feature vector, the feature vector of the first contraction scale corresponding to the second-level feature vector is obtained.

Similarly, the GridNet structure adopted by each level of the network in the above-mentioned network model for extracting the watermark template may be specifically embodied in the manner of obtaining the third-level feature vector according to the second-level feature vector.

First, according to the secondary feature vector, the feature vector of the first shrinkage scale corresponding to the secondary feature vector is obtained. Then, according to the feature vector of the first contraction scale corresponding to the second-level feature vector, the feature vector of the first contraction scale corresponding to the third-level feature vector is obtained. Finally, a three-level feature vector is obtained according to the feature vector of the first shrinkage scale corresponding to the three-level feature vector.

More specifically, according to the feature vector of the first contraction scale corresponding to the second-level feature vector, the feature vector of the first contraction scale corresponding to the third-level feature vector is obtained in the manner described below.

First, according to the feature vector of the first contraction scale corresponding to the second-level feature vector, the feature vector of the second contraction scale corresponding to the second-level feature vector is obtained. Then, according to the feature vector of the second contraction scale corresponding to the second-level feature vector, the feature vector of the second contraction scale corresponding to the third-level feature vector is obtained. Finally, according to the feature vector of the second shrinkage scale corresponding to the three-level feature vector, the feature vector of the first shrinkage scale corresponding to the three-level feature vector is obtained.

At the same time, after obtaining the watermark template, in order to overcome the problem of the existing watermark color diversity, according to the watermark template, the color of the watermark is set to a predetermined color in the carrier object to be processed.

After obtaining the watermark template in step S202, according to the watermark template, the color of the watermark is set to a predetermined color in the carrier object to be processed. Wherein, setting the color of the watermark in the carrier object to be processed to a predetermined color refers to setting the pixel value of the watermark in the carrier object to be processed at the position in the carrier object to be processed to a predetermined Pixel values. In this embodiment, the pixel value of the position of the watermark in the to-be-processed carrier object in the to-be-processed carrier object is set to zero.

Specifically, according to the watermark template, setting the color of the watermark in the carrier object to be processed to a predetermined color may be inputting the carrier object to be processed and the watermark template into a watermark for setting the watermark color. In a network model of a predetermined color, the color of the watermark in the carrier object to be processed is set to a predetermined color.

More specifically, the carrier object to be processed and the watermark template are input into a network model for setting the watermark color to a predetermined color, so as to set the color of the watermark in the carrier object to be processed to a predetermined color, according to The way described below. First, according to the watermark template, the position information and shape information of the watermark in the carrier object to be processed are obtained. Then, according to the position information and shape information of the watermark in the carrier object to be processed, the color of the watermark in the carrier object to be processed is set to a predetermined color.

In the foregoing, the carrier object to be processed and the watermark template are input into the network model for setting the watermark color to a predetermined color, so as to set the color of the watermark in the carrier object to be processed to a predetermined color. It is a network model that sets the watermark color to a predetermined color. Therefore, in this embodiment, it is also necessary to train to obtain a network model for setting the watermark color to a predetermined color. Among them, the network model used to set the watermark color to a predetermined color and the network model used to extract the watermark template belong to mutually corresponding network models.

Obtaining the network model used to set the watermark color to a predetermined color is trained according to the sample data during training of the network model used to extract the watermark template and the watermark template generated during the training process. The specific process of training the network model for setting the watermark color to a predetermined color is described as follows.

First, according to the first-level feature vector and the first-level watermark template, the color of the watermark is set to a predetermined color in the sample data, and the first sample of the sample data set to the predetermined color is obtained. At the same time, according to the secondary feature vector and the secondary watermark template, the color of the watermark is set to a predetermined color in the sample data; the second sample after the sample data is set to the predetermined color is obtained. And, according to the three-level feature vector and the three-level watermark template, the color of the watermark is set to a predetermined color in the sample data; and a third sample of the sample data set to the predetermined color is obtained.

After that, the first sample, the second sample, and the third sample are used as sample data in which the color of the watermark is set to a predetermined color in the sample data.

Finally, the corresponding relationship between the sample data and the sample data of the predetermined color is obtained, and the corresponding relationship is used as a network model for setting the watermark color to the predetermined color.

The main purpose of setting the watermark color to a predetermined color is to unify the colors in the watermarked image. In practice, the watermark colors are diverse, and the diversified watermark colors will cause the watermark to be edited inaccurately during the watermark editing process. Therefore, the watermark color in the image containing the watermark is set to a predetermined color, and the pixel value of the watermark position is set to Zero, the pixel values in other positions remain unchanged. The structure of the network model used to set the watermark color to a predetermined color is similar to the network model used to extract the watermark template. The only difference is the output setting in the structure of the network model used to set the watermark color to a predetermined color The result of the sample data of the predetermined color is three-channel data, and the watermark template output by the network model used to extract the watermark template is two-dimensional single-channel data. Regarding the structure of the network model used to set the watermark color to a predetermined color, refer to the relevant description of the network model used to extract the watermark template for details, which will not be described in detail here.

Setting the pixels of the watermark position to zero can be achieved by using the following formula: y=S^*X, where X is the watermarked image, and S^ is the result of inverting the watermark position of the watermark template, so because S The watermark positions in ^ are all zeros, and the other positions are all ones, so the result obtained is just the result after setting the predetermined color.

In this embodiment, in order to train a more accurate network model for setting the watermark color to a predetermined color and a network model for extracting the watermark template, the sample data for the predetermined color is compared to the watermark template. Correction of the network model; the specific method can be described as follows.

Determine whether the sample data of the predetermined color and the sample data of the desired predetermined color are within the specified threshold; if so, directly use the sample data of the predetermined color as the network model for setting the watermark color to the predetermined color Output the result.

The above-mentioned correction process can be specifically as follows: In the network model used to extract the watermark template, the feature vector generated by each level of the network is denoted as T _i , and the generated sample data of the predetermined color is denoted as M _i . in,

among them,

Indicates the i-th cascade _{network, (F i -1) * (} - 1) simulates the watermark templates bitwise operation, the result of a predetermined color sample data of each stage can be expressed as: M _{_i} = C _i (T _i ), where C _i represents a convolutional layer.

Through the above correction, the two network structures of the network model used to set the watermark color to a predetermined color and the network model used to extract the watermark template can be jointly trained, and have the following advantages: First, the network used to extract the watermark template The watermark template information of the model can provide guidance for the network model used to set the watermark color to a predetermined color. The more accurate the watermark template learning, the more accurate the result of setting the watermark color to the predetermined color; secondly, the watermark color is set to The result of the predetermined color can be effectively fed back to the network model used to extract the watermark template, thereby guiding the generation of the watermark template to a certain extent.

Step S203: Edit the watermark in the carrier object to be processed according to the watermark template to obtain the target carrier object.

After step S202, a watermark template is obtained, and a carrier object in which the color of the watermark in the carrier object to be processed is set to a predetermined color is obtained. Afterwards, the carrier object to be processed, the watermark template, and the carrier object set to a predetermined color are input into the network model for editing the watermark to obtain the target carrier object.

Specifically, the input of the to-be-processed carrier object, the watermark template, and the carrier object set to a predetermined color into the network model for editing the watermark to obtain the target carrier object is as follows.

First, input the carrier object to be processed, the watermark template, and the carrier object set to a predetermined color into the network model for editing watermarks to obtain the carrier object to be processed, and the watermark template And the editing feature vector corresponding to the carrier object set to the predetermined color.

Then, according to the carrier object to be processed, the watermark template, and the editing feature vector corresponding to the carrier object set to a predetermined color, the watermark of the carrier object to be processed is edited to obtain the target carrier object.

Before editing the watermark of the carrier object to be processed to obtain the target carrier object, a network model for editing the watermark needs to be obtained. Wherein, obtaining the network model for editing the watermark may be: first, according to the sample data, the watermark template corresponding to the sample data, and the sample data of the predetermined color, the editing feature vector in the sample data is obtained. Afterwards, according to the editing feature vector, edit the watermark corresponding to the sample data to obtain the target sample data; finally obtain the corresponding relationship between the sample data and the target sample data, and use the corresponding relationship as the target sample data. Network model of watermarking. In this embodiment, it should be noted that editing the watermark can refer to deleting the original watermark, or can also refer to adding the watermark after deleting the original watermark. Correspondingly, the target sample data refers to the original sample data in the network model for generating and editing the watermark or the sample data of the newly added watermark. Of course, in other embodiments of the present application, editing the watermark may also be other watermark editing methods other than detecting the watermark, deleting the original watermark, and adding the watermark, such as modifying the watermark. It still belongs to the protection scope of this application.

The main task of editing a watermark is to accurately restore the watermarked carrier object to the original carrier object without watermark, or add a watermark or modify the watermark after restoring to the original carrier object without watermark. In order to edit the watermark better, the specific position of the watermark in the carrier object must be accurately estimated. Take editing an image with a watermark as an example. Since the color of the watermark in an image with a watermark is uncertain, this brings great challenges to the editing of the watermark. In the structure of the network model used to edit the watermark, the input is not a single watermarked image, and the results obtained in the network model used to extract the watermark template and the network model used to set the watermark color to a predetermined color are also input. Will be entered into the network model used to edit the watermark. For an image, because the network model used to extract the watermark template is a multi-cascade network structure design, the output can be multiple generated watermark templates, and multiple watermark templates are input to the main part of the network model used to edit the watermark. The function is to provide watermark location information and shape information for the watermark editing task; in the network model used to set the watermark color to a predetermined color, multiple images with the watermark set to the predetermined color are obtained, and these images are input for editing the watermark In the network model, the influence of the diversification of watermark colors can be eliminated; in addition, it can also have a certain guiding effect on the position and shape of the watermark. Therefore, in the network model for editing watermarks, multiple images are input, including multiple watermark templates, multiple watermarks are set as images with predetermined colors and a watermark-containing image to be processed. Multiple watermark templates and images with multiple watermarks set to predetermined colors can provide more information for watermark editing to a certain extent, thereby improving the effect of watermark editing. Network Model for editing the structure of the watermark with the above two, the same sampling multi-cascade network structure, in particular, assuming the i-th stage of the network output is expressed as a feature vector V _i, then:

among them

Represents the i-th level repair network,

Is the parameter to be learned.

In addition, since the structure of the network model used to edit the watermark is similar to the network model used to extract the watermark template, the training mode of this network is referred to the training mode of the network model used to extract the watermark template, which will not be repeated here. For the method of adding a new watermark to the carrier after deleting the original watermark after deleting the original watermark, please refer to the relevant description part of the third embodiment, which will not be repeated here.

In this embodiment, a total of three network models are included. Since the network model used to set the watermark color to a predetermined color and the network model used to extract the watermark template are jointly trained, the loss function involved in the network model learning (training) process is mainly divided into two parts. The loss function of the network model used to set the watermark color to a predetermined color and the network model used to extract the watermark template can be expressed as:

among them,

Represents the loss function of the network model used to extract the watermark template,

Represents the loss function of the network model used to set the watermark color to a predetermined color, θ _S represents the learnable parameter in the network model used to extract the watermark template, and θ _T represents the network model used to set the watermark color to the predetermined color Learnable parameters. λ represents a hyperparameter that weighs the two.

Specifically, in the network model used to extract the watermark template, given a watermarked image x, it is expected that different multi-cascade sub-networks can output the watermark template S, and the loss function can be expressed as:

Among them, L is the number of multiple cascades in the network model used to extract the watermark template, C _i represents a convolutional layer, and the i-th level network of the network model used to extract the watermark template will output a feature vector F _i , This feature vector will provide guidance for the network model used to set the watermark color to a predetermined color, and its loss function can be expressed as:

Among them, T is the real result after the watermark color is set to the predetermined color, I(F _i )=(F _i -1)*(-1), which means the operation of simulating bitwise inversion.

In the network model used to edit the watermark, its network structure also adopts a multi-cascade network structure. Assuming that the real image without watermark is represented as y, the loss function of the network model used to edit the watermark can be expressed as:

among them,

Represents a set, including watermarked images, and multiple watermark templates obtained from the network model used to extract the watermark template and multiple watermarks obtained from the network model used to set the watermark color to a predetermined color. image.

Please refer to FIG. 3, which shows a schematic diagram of a second application scenario embodiment of the carrier object processing method provided by the present application. The user inputs the image with the watermark "2018" and the editing requirement "modify the watermark of the image 2018 to 2019" into the computing device 300, and the computing device 300 obtains the specific position of the watermark "2008" in the image by extracting the watermark template of the image. After that, set the watermark "2018" to a predetermined color according to the specific location of the watermark, and set the watermark to a specific color to overcome the inaccuracy of the watermark edited during the watermark editing process caused by the diversification of the watermark color, and finally set it to The watermark "2018" of a specific color is changed to the watermark "2019". Of course, the above-mentioned image with the watermark "2018" can also be input alone or together with other batches of images. This method can automatically identify the image with the watermark "2018" and automatically detect its watermark "2018" according to the actual time. , And automatically modify the watermark time "2018" in the image according to the current time. In the process of specific watermark detection and modification, multiple sets of watermark modification schemes can be automatically recommended based on the results of the detection and recognition of the watermark. When modifying the watermark, you can choose one of the recommended multiple sets of modification watermark schemes for modification, or it can be Modified into multiple. For example, the watermark in the image with the watermark "2018" can be modified to "2019", or it can be modified to "2019" at the same time. Of course, it is also possible to recognize only the watermark in the image and save the recognized watermark result for subsequent modification. The above-mentioned image containing the watermark "2018" can be automatically modified according to the current time after it has been modified. For example, the watermark "2018" of the image can be changed to "2019" in 2019, and it can be automatically changed to "2020" in 2020. Of course, the purpose of the foregoing application scenario embodiments is to facilitate the understanding of the carrier object processing method of the present application, and is not used to limit the carrier object processing method of the present application.

The carrier object processing method of this embodiment can accurately obtain the specific position of the watermark by extracting the watermark template corresponding to the carrier object to be processed; according to the watermark template, the watermark of the carrier object to be processed can be accurately edited. The carrier object processing method of the present application solves the problem that the existing watermark editing method cannot accurately edit the watermark due to the diversity of watermark styles by obtaining the watermark template corresponding to the carrier object to be processed.

In the first embodiment described above, a carrier object processing method is provided, and correspondingly, the present application also provides a carrier object processing device. As shown in FIG. 5, it is a schematic diagram of a carrier object processing apparatus according to the second embodiment of the application. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment. The device embodiments described below are merely illustrative.

A carrier object processing device of this embodiment includes:

The carrier object processing unit 501 is configured to obtain a carrier object to be processed;

A watermark template extraction unit 502, configured to extract a watermark template from the carrier object to be processed;

The target carrier object obtaining unit 503 is configured to edit the watermark in the to-be-processed carrier object according to the watermark template to obtain a target carrier object.

Optionally, the watermark template extraction unit is specifically used for:

Inputting the to-be-processed carrier object into a network model for extracting a watermark template to obtain a feature vector of the to-be-processed carrier object;

Optionally, the watermark template extraction unit is specifically used for:

Optionally, it also includes: a color setting unit;

The color setting unit is specifically configured to: set the color of the watermark to a predetermined color in the carrier object to be processed according to the watermark template;

The target carrier object obtaining unit is specifically used for:

According to the predetermined color of the watermark and the watermark template, edit the watermark in the to-be-processed carrier object to obtain a target carrier object.

Optionally, the color setting unit is specifically used for:

Optionally, it further includes: a second carrier object obtaining unit;

The second carrier object obtaining unit is specifically used for:

The target carrier object obtaining unit is specifically used for:

The carrier object to be processed, the watermark template, and the carrier object set to a predetermined color are input into a network model for editing the watermark to obtain a target carrier object.

Optionally, the target carrier object obtaining unit is specifically used for:

According to the to-be-processed carrier object, the watermark template, and the editing feature vector corresponding to the carrier object set to a predetermined color, edit the watermark of the to-be-processed carrier object to obtain the target carrier object.

Optionally, it further includes a first network model obtaining unit;

The first network model obtaining unit is configured to obtain the network model used to extract the watermark template;

The first network model obtaining unit is specifically used for:

Obtain sample data for extracting the watermark template;

Obtaining the first-level feature vector of the sample data;

Optionally, the first network model obtaining unit is specifically configured to:

Optionally, it further includes a second network model obtaining unit;

The second network model obtaining unit is configured to obtain a network model for setting the watermark color to a predetermined color;

The second network model obtaining unit is specifically used for:

Optionally, it also includes a correction unit;

The correction unit is configured to correct the network model used for extracting the watermark template according to the sample data of the predetermined color;

The correction unit is specifically used for:

Optionally, it further includes a third network model obtaining unit;

The third network model obtaining unit is configured to obtain a network model for editing watermarks;

The third network model obtaining unit is specifically used for:

Optionally, the watermark template extraction unit is specifically used for:

This application also provides a watermark embedding method, as shown in FIG. 6, which is a flowchart of an embodiment of the watermark embedding method in the third embodiment of this application. The method includes the following steps. include:

Step S601: Obtain the original carrier object, and obtain the original watermark.

The watermark embedding method of this embodiment can be used to generate a watermark-containing carrier object to train the network model for extracting the watermark template of the first embodiment, the network model for setting the watermark color to a predetermined color, and the network for editing the watermark model.

Specifically, the original watermark can be obtained in the manner described below.

First, randomly combine at least one of numeric symbols, English symbols, and special symbols to form a character sequence. After that, the character sequence is used as the original watermark.

For example, A number symbols, B English symbols, and C special symbols can be selected in advance to randomly form a sequence containing A+B+C characters, and the character sequence is used as the original watermark.

Step S602: Obtain the position information at which the original watermark is embedded on the original carrier object.

After obtaining the original carrier object and obtaining the original watermark in step S601, the position information at which the original watermark is embedded on the original carrier object is obtained.

There are many ways to obtain the position information of the original watermark embedded on the original carrier object, and one of the ways to obtain it may be: firstly, obtain the size information of the original carrier object and the information of the original watermark. Size Information. Then, according to the size information of the original carrier object, the embedding area of the original watermark on the original carrier object is determined. Finally, in the embedding area, according to the size information of the original watermark, the position information at which the original watermark is embedded on the original carrier object is obtained.

Further, the determination of the embedding area of the original watermark on the original carrier object according to the size information of the original carrier object may be in the manner described below. First, obtain the length and size information of the original carrier object according to the size information of the original carrier object. After that, determine the first reserved size information in the length direction on the original carrier object according to the length and size information; at the same time, obtain the width and size information of the original carrier object according to the size information of the original carrier object; And, determining the second reserved size information in the width direction on the original carrier object according to the width size information. Finally, according to the size information of the original carrier object, the first reserved size information and the second reserved size information, the embedding area of the original watermark on the original carrier object is determined.

More specifically, in the embedding area, according to the size information of the original watermark, the position information at which the original watermark is embedded on the original carrier object is obtained. First, according to the size information of the original watermark, Obtain the edge position information of the original watermark. Then, according to the edge position information of the original watermark and the position information of the embedding area, the position information of the original watermark embedded on the original carrier object is obtained.

Step S603: In the original carrier object, randomly according to the attribute information of the original watermark and the position information, embed the original watermark in the designated position of the original carrier object.

In step S601, the original carrier object is obtained, and the original watermark is obtained; and in step S602, after obtaining the position information of the original watermark embedded on the original carrier object, in the original carrier object, randomly follow the attributes of the original watermark Information and the location information, the original watermark is embedded in the specified location of the original carrier object.

Wherein, the attribute information of the original watermark includes at least one of the color of the original watermark, the font of the original watermark, and the font size information of the original watermark.

Using the above to obtain the position information of the original watermark embedded on the original carrier object, in order to prevent the watermark from being on the edge of the original carrier object, a random position generation rule can be adopted: X=random(W/10, WW/10), Y=random(H/10, HH/10) generates the position information of the original watermark, where X and Y are the coordinate positions of the watermark on the original carrier object, and W and H are the width and height of the original carrier object.

The watermark embedding method of this embodiment obtains the original carrier object and the original watermark. On this basis, the position information of the original watermark embedded in the original carrier object is obtained, so that in the original carrier object, randomly according to the attribute information of the original watermark and the position information, the original The watermark is embedded in the designated position of the original carrier object. The watermark embedded in this method ensures that the watermark is not on the edge of the original carrier object, so as to ensure that the watermark can be accurately detected.

In the above-mentioned third embodiment, a watermark embedding method is provided. Correspondingly, this application also provides a watermark embedding device. As shown in FIG. 7, it is a schematic diagram of a watermark embedding device according to the fourth embodiment of the application. Since the device embodiment is basically similar to the method embodiment, the description is relatively simple, and for related parts, please refer to the part of the description of the method embodiment. The device embodiments described below are merely illustrative.

A watermark embedding device of this embodiment includes:

The original carrier object obtaining unit 701 is configured to obtain the original carrier object;

The original watermark obtaining unit 702 is configured to obtain the original watermark;

A location information obtaining unit 703, configured to obtain location information of the original watermark embedded on the original carrier object;

The embedding unit 704 is configured to embed the original watermark at a designated position of the original carrier object randomly according to the attribute information of the original watermark and the position information in the original carrier object.

Optionally, the original watermark obtaining unit is specifically used for:

Use the character sequence as the original watermark.

Optionally, the location information obtaining unit is specifically configured to:

The first embodiment of the present application provides a method for processing a carrier object, and the fifth embodiment of the present application provides an electronic device corresponding to the method for processing the carrier object.

As shown in FIG. 8, it shows a schematic diagram of the electronic device of the carrier object processing method provided in the fifth embodiment of the present application.

An electronic device of this embodiment includes:

Processor 801;

The memory 802 is used to store the program of the carrier object processing method and execute the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

The third embodiment of this application provides a watermark embedding method, and the sixth embodiment of this application provides an electronic device corresponding to the watermark embedding method. Since the schematic diagram of the electronic device is the same as FIG. 8, please refer to FIG. 8 for the specific diagram.

This application provides an electronic device, including:

processor;

Obtain the original carrier object and obtain the original watermark;

The first embodiment of the present application provides a method for processing a carrier object, and the seventh embodiment of the present application provides a computer storage medium corresponding to the method for processing the carrier object.

A computer storage medium of this embodiment, the computer storage medium stores a program of a carrier object processing method, and the program is run by a processor to perform the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

The third embodiment of the present application provides a watermark embedding method, and the eighth embodiment of the present application provides a computer storage medium corresponding to the watermark embedding method.

A computer storage medium of this embodiment, the computer storage medium stores a program of the watermark embedding method, and the program is run by a processor to perform the following steps:

Obtain the original carrier object and obtain the original watermark;

Although this application is disclosed as above in preferred embodiments, it is not intended to limit this application. Any person skilled in the art can make possible changes and modifications without departing from the spirit and scope of this application. Therefore, this application The scope of protection shall be subject to the scope defined by the claims of this application.

In a typical configuration, the computing device includes one or more processors (CPUs), input/output interfaces, network interfaces, and memory.

The memory may include non-permanent memory in a computer readable medium, random access memory (RAM) and/or non-volatile memory, such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer readable media.

Computer-readable media include permanent and non-permanent, removable and non-removable media, and information storage can be realized by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technology, CD-ROM, digital versatile disc (DVD) or other optical storage, Magnetic cassettes, magnetic tape magnetic disk storage or other magnetic storage devices or any other non-transmission media can be used to store information that can be accessed by computing devices. According to the definition in this article, computer-readable media does not include non-transitory computer-readable media (transitory media), such as modulated data signals and carrier waves.

Those skilled in the art should understand that the embodiments of the present application can be provided as methods, systems, or computer program products. Therefore, this application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, this application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.

Claims

A method for processing a carrier object, characterized in that it comprises:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

According to the watermark template, edit the watermark in the to-be-processed carrier object to obtain a target carrier object.
The method according to claim 1, wherein the extracting a watermark template from the carrier object to be processed comprises:

Input the to-be-processed carrier object into a network model for extracting a watermark template to obtain a feature vector of the to-be-processed carrier object;

According to the feature vector of the carrier object to be processed, a watermark template corresponding to the feature vector is obtained.
The method according to claim 2, wherein the obtaining the watermark template corresponding to the feature vector according to the feature vector of the carrier object to be processed comprises:

Performing a convolution operation on the feature vector to obtain a watermark template corresponding to the feature vector.
The method according to claim 1, wherein the watermark template is a template used when adding watermark information to the original carrier object; wherein the original carrier object is the carrier before the watermark information is added to the carrier object to be processed Object.
The method according to claim 1, further comprising:

Setting the color of the watermark in the carrier object to be processed to a predetermined color according to the watermark template;

The editing the watermark in the carrier object to be processed according to the watermark template to obtain a target carrier object includes: editing in the carrier object to be processed according to a predetermined color of the watermark and the watermark template The watermark obtains the target carrier object.
The method according to claim 5, wherein the setting the color of the watermark in the carrier object to be processed to a predetermined color according to the watermark template comprises:

The carrier object to be processed and the watermark template are input into a network model for setting the watermark color to a predetermined color, so as to set the color of the watermark in the carrier object to be processed to a predetermined color.
The method according to claim 6, wherein the input of the carrier object to be processed and the watermark template into a network model for setting the watermark color to a predetermined color, so as to transfer the carrier object to be processed The color of the watermark in the object is set to a predetermined color, including:

Obtaining the position information and shape information of the watermark in the carrier object to be processed according to the watermark template;

According to the position information and shape information of the watermark in the carrier object to be processed, the color of the watermark in the carrier object to be processed is set to a predetermined color.
The method according to claim 5, further comprising:

Obtaining the carrier object in which the color of the watermark in the carrier object to be processed is set to a predetermined color;

The editing the watermark in the to-be-processed carrier object according to the predetermined color of the watermark and the watermark template to obtain a target carrier object includes:

The carrier object to be processed, the watermark template, and the carrier object set to a predetermined color are input into a network model for editing the watermark to obtain a target carrier object.
The method according to claim 8, wherein said inputting said carrier object to be processed, said watermark template and said carrier object set to a predetermined color into a network model for editing watermark, Get the target carrier object, including:

Input the to-be-processed carrier object, the watermark template, and the carrier object set to a predetermined color into the network model for editing watermarks to obtain the to-be-processed carrier object, the watermark template, and the The edit feature vector corresponding to the carrier object set as a predetermined color;

According to the to-be-processed carrier object, the watermark template, and the editing feature vector corresponding to the carrier object set to a predetermined color, edit the watermark of the to-be-processed carrier object to obtain a target carrier object.
The method according to claim 5, wherein the setting the color of the watermark in the carrier object to be processed to a predetermined color refers to setting the watermark in the carrier object to be processed on the carrier object to be processed. The pixel value at the position in the object is set to a predetermined pixel value.
The method according to claim 2, further comprising obtaining the network model used to extract the watermark template;

The obtaining of the network model used to extract the watermark template includes:

Obtain sample data for extracting the watermark template;

Obtaining the first-level feature vector of the sample data;

Performing a convolution operation on the first-level feature vector to obtain a first-level watermark template corresponding to the first-level feature vector;

Obtaining a second-level feature vector according to the first-level feature vector;

Performing a convolution operation on the secondary feature vector to obtain a secondary watermark template corresponding to the secondary feature vector;

Obtaining a third-level feature vector according to the second-level feature vector;

Performing a convolution operation on the three-level feature vector to obtain a three-level watermark template corresponding to the three-level feature vector;

Using the first-level watermark template, the second-level watermark template, and the third-level watermark template as watermark templates corresponding to the sample data;

The corresponding relationship between the sample data and the watermark template corresponding to the sample data is obtained, and the corresponding relationship is used as a network model for extracting the watermark template.
The method according to claim 11, wherein the obtaining a second-level feature vector according to the first-level feature vector comprises:

Obtaining, according to the first-level feature vector, the feature vector of the first shrinkage scale corresponding to the first-level feature vector;

Obtaining the feature vector of the first contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector;

The second-level feature vector is obtained according to the feature vector of the first shrinkage scale corresponding to the second-level feature vector.
The method according to claim 12, wherein the obtaining the feature vector of the first contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector comprises :

Obtaining the feature vector of the second contraction scale corresponding to the first-level feature vector according to the feature vector of the first contraction scale corresponding to the first-level feature vector;

Obtaining the feature vector of the second contraction scale corresponding to the second-level feature vector according to the feature vector of the second contraction scale corresponding to the first-level feature vector;

According to the feature vector of the second contraction scale corresponding to the second-level feature vector, the feature vector of the first contraction scale corresponding to the second-level feature vector is obtained.
The method according to claim 11, wherein the obtaining a third-level feature vector according to the second-level feature vector comprises:

Obtaining, according to the second-level feature vector, a feature vector of the first shrinkage scale corresponding to the second-level feature vector;

Obtaining the feature vector of the first contraction scale corresponding to the third-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector;

According to the feature vector of the first shrinkage scale corresponding to the three-level feature vector, a three-level feature vector is obtained.
The method according to claim 14, wherein the obtaining the feature vector of the first contraction scale corresponding to the third-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector comprises :

Obtaining the feature vector of the second contraction scale corresponding to the second-level feature vector according to the feature vector of the first contraction scale corresponding to the second-level feature vector;

Obtaining the feature vector of the second contraction scale corresponding to the third-level feature vector according to the feature vector of the second contraction scale corresponding to the second-level feature vector;

According to the feature vector of the second contraction scale corresponding to the three-level feature vector, the feature vector of the first contraction scale corresponding to the three-level feature vector is obtained.
The method according to claim 11, further comprising obtaining a network model for setting the watermark color to a predetermined color;

The obtaining a network model for setting the watermark color to a predetermined color includes:

According to the first-level feature vector and the first-level watermark template, set the color of the watermark in the sample data to a predetermined color, and obtain a first sample of the sample data set to the predetermined color;

Setting the color of the watermark in the sample data to a predetermined color according to the secondary feature vector and the secondary watermark template; obtaining a second sample of the sample data set to the predetermined color;

According to the three-level feature vector and the three-level watermark template, set the color of the watermark in the sample data to a predetermined color; obtain a third sample of the sample data set to the predetermined color;

Using the first sample, the second sample, and the third sample as sample data in which the color of the watermark is set to a predetermined color in the sample data;

The corresponding relationship between the sample data and the sample data of the predetermined color is obtained, and the corresponding relationship is used as a network model for setting the watermark color to the predetermined color.
The method according to claim 16, further comprising: correcting the network model for extracting the watermark template according to the sample data of the predetermined color;

The correcting the network model for extracting the watermark template according to the sample data of the predetermined color includes:

Determine whether the sample data of the predetermined color and the sample data of the desired predetermined color are within the specified threshold; if so, directly use the sample data of the predetermined color as the network model for setting the watermark color to the predetermined color Output result;

If not, reacquire the correction feature vector of the carrier object to be processed according to the sample data of the predetermined color and the sample data used to extract the watermark template; obtain the correction feature vector according to the correction feature vector Corresponding corrected watermark template.
The method according to claim 16, further comprising obtaining a network model for editing the watermark;

The obtaining of the network model used for editing the watermark includes:

Obtaining the edit feature vector in the sample data according to the sample data, the watermark template corresponding to the sample data, and the sample data of the predetermined color;

According to the editing feature vector, edit the watermark corresponding to the sample data to obtain target sample data;

The corresponding relationship between the sample data and the target sample data is obtained, and the corresponding relationship is used as a network model for editing the watermark.
The method according to claim 1, wherein the extracting a watermark template from the carrier object to be processed comprises:

Automatically obtain at least one set of watermark templates corresponding to the to-be-processed carrier object according to the to-be-processed carrier object;

A set of watermark templates or all of the watermark templates in the at least one set of watermark templates corresponding to the carrier object to be processed are used as the watermark templates of the carrier to be processed.
A carrier object processing device, characterized in that it comprises:

The carrier object processing unit is used to obtain the carrier object to be processed;

A watermark template extraction unit, configured to extract a watermark template from the carrier object to be processed;

The target carrier object obtaining unit is configured to edit the watermark in the to-be-processed carrier object according to the watermark template to obtain the target carrier object.
A watermark embedding method is characterized in that it includes:

Obtain the original carrier object and obtain the original watermark;

Obtaining position information at which the original watermark is embedded on the original carrier object;

In the original carrier object, randomly according to the attribute information of the original watermark and the position information, the original watermark is embedded in a designated position of the original carrier object.
The method according to claim 21, wherein said obtaining the original watermark comprises:

Randomly combine at least one of numeric symbols, English symbols, and special symbols to form a character sequence;

Use the character sequence as the original watermark.
The method according to claim 21, wherein said obtaining the position information of the original watermark embedded on the original carrier object comprises:

Acquiring the size information of the original carrier object and the size information of the original watermark;

Determine the embedding area of the original watermark on the original carrier object according to the size information of the original carrier object;

In the embedding area, according to the size information of the original watermark, the position information at which the original watermark is embedded on the original carrier object is obtained.
The method according to claim 23, wherein the determining the embedding area of the original watermark on the original carrier object according to the size information of the original carrier object comprises:

Obtaining the length and size information of the original carrier object according to the size information of the original carrier object;

Determine the first reserved size information in the length direction on the original carrier object according to the length size information;

Obtaining the width size information of the original carrier object according to the size information of the original carrier object;

Determine the second reserved size information in the width direction on the original carrier object according to the width size information;

According to the size information of the original carrier object, the first reserved size information and the second reserved size information, the embedding area of the original watermark on the original carrier object is determined.
The method according to claim 24, wherein the obtaining, in the embedding area, according to the size information of the original watermark, the position information at which the original watermark is embedded on the original carrier object comprises:

Obtaining edge position information of the original watermark according to the size information of the original watermark;

According to the edge position information of the original watermark and the position information of the embedding area, the position information at which the original watermark is embedded on the original carrier object is obtained.
The method according to claim 22, wherein the attribute information of the original watermark includes at least one of the color of the original watermark, the font of the original watermark, and the font size information of the original watermark.
A watermark embedding device is characterized in that it comprises:

The original carrier object obtaining unit, which is used to obtain the original carrier object;

The original watermark obtaining unit is used to obtain the original watermark;

A location information obtaining unit, configured to obtain location information of the original watermark embedded on the original carrier object;

The embedding unit is configured to embed the original watermark at a designated position of the original carrier object randomly according to the attribute information of the original watermark and the position information in the original carrier object.
An electronic device, characterized in that it comprises:

processor;

The memory is used to store the program of the carrier object processing method and execute the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

According to the watermark template, edit the watermark in the to-be-processed carrier object to obtain a target carrier object.
An electronic device, characterized in that it comprises:

processor;

The memory is used to store the program of the watermark embedding method and execute the following steps:

Obtain the original carrier object and obtain the original watermark;

Obtaining position information at which the original watermark is embedded on the original carrier object;

In the original carrier object, randomly according to the attribute information of the original watermark and the position information, the original watermark is embedded in a designated position of the original carrier object.
A computer storage medium, characterized in that the computer storage medium stores a program of a carrier object processing method, and the program is run by a processor to perform the following steps:

Obtain the to-be-processed carrier object;

Extracting a watermark template from the to-be-processed carrier object;

According to the watermark template, edit the watermark in the to-be-processed carrier object to obtain a target carrier object.
A computer storage medium, characterized in that the computer storage medium stores a program of a watermark embedding method, and the program is run by a processor to perform the following steps:

Obtain the original carrier object and obtain the original watermark;

Obtaining position information at which the original watermark is embedded on the original carrier object;

In the original carrier object, randomly according to the attribute information of the original watermark and the position information, the original watermark is embedded in a designated position of the original carrier object.