CN112613045B - Method and system for embedding data watermark of target data - Google Patents

Abstract

The invention discloses a data watermark embedding method and system for target data. The method includes: S1 dividing the target data to be embedded with a data watermark into a plurality of content blocks, and embedding a data watermark in each content block; S2 adopting a preset The data similarity evaluation model evaluates the data item similarity of the data items embedded in the data watermark; S3 evaluates the similarity of the content block data watermark based on the similarity of all data items that make up the content block. When the data watermarks of each content block are similar Execute S4 when the degrees all meet the first threshold range, otherwise adjust the embedding ratio and/or position of the data watermark in the content block and execute S2; S4 calculate the overall similarity of the target data based on the data watermark similarities of all content blocks that make up the target data The target data embedded in the data watermark is obtained by adjusting the embedding ratio and/or position of the data watermark. Finally, the high concealment and high emulation of data watermark embedding are realized.

Description

A method and system for embedding data watermark of target data

Technical Field

The present invention relates to the field of data watermarking, and in particular to a method and system for embedding data watermarking into target data.

Background Art

With the continuous development of the digital economy, information exchange between different departments, different regions, and different data subjects has gradually increased, and data has been circulated, reorganized, and used more and more frequently in the form of structured data between various links. The use of data in a dynamic environment poses a huge risk of data leakage. Once a data leakage occurs, it is necessary to accurately locate the responsible link to trace the security responsibility of the relevant personnel and strengthen the security control of the weak links in a targeted manner.

Data watermark technology is one of the effective technical means to solve the above-mentioned responsibility tracing problem after data leakage. Data watermarking is to add additional redundant identification information to the data content itself. It can associate and record the relevant responsibility links by highly simulating the data content and incorporating identification information. Once the data is leaked, it can be located according to the watermark information added in advance. High simulation and high concealment are the key indicators for the effectiveness of data watermarking to avoid being discovered and destroyed by malicious users. The realization of high simulation and high concealment of data watermarking requires that the similarity of the target data before and after the data watermark is added must reach a threshold that is not easy for users to find. Therefore, how to embed data watermarks in the target data to achieve high concealment and high simulation after the data watermark is embedded needs to be solved urgently.

Summary of the invention

In order to solve the above-mentioned deficiencies in the prior art, the present invention provides a data watermark embedding method for target data, comprising:

S1 divides the target data to be embedded with a data watermark into a plurality of content blocks, and embeds a data watermark in a data entry of each content block;

S2 uses a preset data similarity evaluation model to evaluate the data entry similarity of the data entry embedded with the data watermark;

S3 evaluates the similarity of data watermarks of the content blocks based on the data entry similarities of all data entries constituting the content blocks. When the similarities of the data watermarks of the content blocks all meet the first threshold range, execute S4; otherwise, adjust the embedding ratio and/or position of the data watermarks in the content blocks that do not meet the first threshold range, and execute S2;

S4 calculates the similarity of the target data as a whole based on the similarity of the data watermarks of all the content blocks constituting the target data. When the similarity of the target data as a whole satisfies a second threshold range, the embedding of the data watermark is completed. Otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and S2 is executed.

Preferably, adjusting the embedding ratio and/or position of the data watermark includes:

When the data entry contains a single type of field, adjust the embedding ratio of the data watermark;

When a data entry contains multiple types of fields, the embedding ratio and/or position of the data watermark is adjusted.

Preferably, the adjusting the embedding ratio of the data watermark includes:

When the content block data watermark similarity is greater than the maximum value in the first threshold range, reducing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

When the content block data watermark similarity is less than the minimum value in the first threshold range, increasing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

When the overall similarity of the target data is greater than the maximum value in the second threshold range, reducing the proportion of the data watermark embedded in one or more content blocks to a preset proportion;

When the overall similarity of the target data is less than the minimum value in the second threshold range, the proportion of the data watermark embedded in one or more content blocks is increased to a preset proportion.

Preferably, the step of adjusting the embedding position of the data watermark includes:

Remove the original data watermark in the data entry, and embed data watermarks matching the field type into various types of fields in the data entry according to a preset ratio;

The data entries after embedding the data watermark matching the field type are evaluated for data entry similarity, the position of the field with the greatest data entry similarity is selected as the optimal position for embedding the data watermark, and the data watermark is embedded at the optimal position.

Preferably, the field types in the data entry include any one or more of the following:

Numeric fields, text fields, and natural language fields.

Preferably, embedding a data watermark in the data entry comprises:

When the data entry includes a numerical field, embedding a numerical data watermark in the numerical field;

When the data entry includes a text field, embedding a data watermark of a character text type in the text field;

When the data entry includes a natural language field, a natural language data watermark is embedded in the natural language field.

Preferably, the use of a preset data similarity evaluation model to perform data entry similarity evaluation on the data entries embedded with the data watermark includes:

When a numerical data watermark is embedded in the numerical field of the data entry, the numerical values before and after the data watermark is embedded are deconstructed and segmented, and the data entry similarity is evaluated using a Euclidean distance vector data similarity evaluation model;

When a character text type data watermark is embedded in the text field of the data entry, the ASCII code values before and after the data watermark is embedded are deconstructed, and the data entry similarity is evaluated by a cosine vector data similarity evaluation model;

When a natural language data watermark is embedded in the natural language field of the data entry, a space vector model is applied to the natural language field before and after the data watermark is embedded to perform word deconstruction and segmentation, and the deconstruction and segmentation results are evaluated for data entry similarity using a cosine vector data similarity evaluation model.

Preferably, the content block data watermark similarity evaluation is performed according to the following formula:

Where: δ represents the content block data watermark similarity; N represents the total number of data entries in the content block; _Ci represents the data entry similarity of the i-th data entry.

Preferably, the similarity of the target data as a whole is evaluated as follows:

Where: θ represents the similarity of the target data as a whole; M represents the total number of content blocks in the target data; _δi represents the watermark similarity of the content block data of the i-th content block.

Based on the same inventive concept, the present invention also provides a data watermark embedding system for target data, comprising:

An embedding module, used for dividing the target data to be embedded with a data watermark into a plurality of content blocks, and embedding a data watermark in each content block;

A data entry similarity evaluation module is used to use a preset data similarity evaluation model to perform data entry similarity evaluation on the data entry after the data watermark is embedded;

A content block similarity evaluation module, used to evaluate the similarity of data watermarks of the content blocks based on the data entry similarities of all data entries constituting the content blocks, and when the similarities of the data watermarks of the content blocks all meet the first threshold range, the overall similarity evaluation module is executed, otherwise, the embedding ratio and/or position of the data watermarks in the content blocks that do not meet the first threshold range is adjusted, and the data entry similarity evaluation module is executed;

The overall similarity evaluation module is used to calculate the overall similarity of the target data based on the data watermark similarity of all content blocks that constitute the target data. When the overall similarity of the target data meets the second threshold range, the embedding of the data watermark is completed; otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and the data entry similarity evaluation module is executed.

Preferably, the data entry similarity assessment module is specifically used to:

When a numerical data watermark is embedded in the numerical field of the data entry, the numerical values before and after the data watermark is embedded are deconstructed and segmented, and the data entry similarity is evaluated using a Euclidean distance vector data similarity evaluation model;

When a character text type data watermark is embedded in the text field of the data entry, the ASCII code values before and after the data watermark is embedded are deconstructed, and the data entry similarity is evaluated by a cosine vector data similarity evaluation model;

When a natural language data watermark is embedded in the natural language field of the data entry, a space vector model is applied to the natural language field before and after the data watermark is embedded to perform word deconstruction and segmentation, and the deconstruction and segmentation results are evaluated for data entry similarity using a cosine vector data similarity evaluation model.

Compared with the prior art, the present invention has the following beneficial effects:

The technical solution provided by the present invention is as follows: S1 divides the target data to be embedded with a data watermark into multiple content blocks, and embeds a data watermark in the data entry of each content block; S2 uses a pre-set data similarity evaluation model to perform data entry similarity evaluation on the data entry after the data watermark is embedded; S3 performs content block data watermark similarity evaluation based on the data entry similarity of all data entries constituting the content block, and when the data watermark similarity of each content block meets the first threshold range, S4 is executed, otherwise the embedding ratio and/or position of the data watermark in the content block that does not meet the first threshold range is adjusted, and S2 is executed; S4 calculates the similarity of the target data as a whole based on the data watermark similarity of all content blocks constituting the target data, and when the similarity of the target data as a whole meets the second threshold range, the embedding of the data watermark is completed, otherwise the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and S2 is executed. The present invention dynamically adjusts the data watermark embedded in the content block according to the data watermark similarity evaluation results of the data entry, the content block and the data as a whole, so as to finally achieve high concealment and high simulation after the data watermark is embedded.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG1 is a flow chart of a method for embedding a data watermark into target data provided by the present invention;

FIG. 2 is a schematic diagram of a data watermark embedding system for target data provided by an embodiment of the present invention.

DETAILED DESCRIPTION

In order to better understand the present invention, the content of the present invention is further described below in conjunction with the accompanying drawings and examples.

Embodiment 1: As shown in FIG1 , in order to meet the urgent needs of the above-mentioned prior art, the present invention provides a data watermark embedding method for target data, comprising:

S1 divides the target data to be embedded with a data watermark into a plurality of content blocks, and embeds a data watermark in a data entry of each content block;

S2 uses a preset data similarity evaluation model to evaluate the data entry similarity of the data entry embedded with the data watermark;

S3 evaluates the similarity of data watermarks of the content blocks based on the data entry similarities of all data entries constituting the content blocks. When the similarities of the data watermarks of the content blocks all meet the first threshold range, execute S4; otherwise, adjust the embedding ratio and/or position of the data watermarks in the content blocks that do not meet the first threshold range, and execute S2;

S4 calculates the similarity of the target data as a whole based on the similarity of the data watermarks of all the content blocks constituting the target data. When the similarity of the target data as a whole satisfies a second threshold range, the embedding of the data watermark is completed. Otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and S2 is executed.

Wherein, adjusting the embedding ratio and/or position of the data watermark includes:

When the data entry contains a single type of field, adjust the embedding ratio of the data watermark;

When a data entry contains multiple types of fields, the embedding ratio and/or position of the data watermark is adjusted.

The present invention sequentially evaluates the data watermark similarity of data entries, content blocks and the data as a whole, and dynamically adjusts the data watermark embedded in the target data so that the similarity of the content block after the data watermark is embedded and the similarity of the target data as a whole meet the set threshold ranges respectively, so as to finally achieve high concealment and high simulation after the data watermark is embedded.

In this embodiment, S1 divides the target data to be embedded with a data watermark into multiple content blocks, and embeds a data watermark in a data entry of each content block, including:

For each data entry constituting the content block, a data watermark of a corresponding type is selected and embedded according to the field type in the data entry. In order to increase the information capacity of the embedded data watermark, the proportion of the data watermark embedded in the target data is 100%.

Specifically include:

When the data entry includes a numerical field, embedding a numerical data watermark in the numerical field;

When the data entry includes a text field, embedding a data watermark of a character text type in the text field;

When the data entry includes a natural language field, a natural language data watermark is embedded in the natural language field.

S2 uses a preset data similarity evaluation model to perform data entry similarity evaluation on the data entries embedded with the data watermark, that is, according to different data watermark embedding algorithms, selects a suitable data similarity evaluation model to perform similarity evaluation on the data watermark entries, including:

The similarity in this embodiment means that for a certain type of data, after the data watermark is embedded, its data type characteristics should not change. If it changes, the similarity evaluation result of its data watermark entry is 0.

For example, for data of the mobile phone number type, the existing mobile phone number is 11 digits, of which the first 3 digits represent the network identification number, the 4th to 7th digits represent the area code, and the 8th to 11th digits represent the user number. After the data watermark is embedded, it should still conform to the data type characteristics of the mobile phone number.

(1) After embedding the numerical data watermark, the numerical values before and after the data watermark is embedded should be deconstructed and segmented, and the similarity evaluation should be performed using the Euclidean distance vector data similarity evaluation model. The evaluation result is D.

For example, for data of the mobile phone number type, the values before and after the data watermark is embedded are P and P' respectively. After structural segmentation, each digit should be an independent unit, that is, P = {N1, N2, ..., N11}, P' = {N'1, N'2, ..., N'11}; then bring it into the Euclidean data similarity evaluation model to calculate the similarity

(2) After embedding the character text data watermark, the ASCII code values before and after the data watermark is embedded should be deconstructed, and the similarity evaluation should be performed using the cosine vector data similarity evaluation model. The evaluation result is C.

For example, for WeChat account type data, the values before and after data watermark embedding are Pi and Pi' respectively. After ASCII code value deconstruction, each digit should be an independent unit, that is, P = {N1, N2, ..., Nn}, P' = {N'1, N'2, ..., N'n}; then bring it into the cosine data similarity evaluation model to calculate the similarity

(3) After the natural language type data watermark is embedded, the space vector model is used to deconstruct and segment the data before and after the data watermark is embedded, and the data similarity of the deconstructed segmentation results is evaluated using the cosine vector data similarity evaluation model.

The natural language data involved in the power business has significant professional characteristics, such as address data such as maintenance address and expansion address; power professional terminology data such as operation terminology and electrical quantity terminology; names of residents, etc., which can form a characteristic word segmentation library of natural language data for the power business.

The natural language data of the power business before and after adding the data watermark is processed by word segmentation, and the vector expressions obtained are O = {O1, O2, ..., On} and O' = {O'1, O'2, ..., O'n}, which are brought into the cosine data similarity evaluation model to calculate the similarity

S3 evaluates the similarity of data watermarks of the content blocks based on the data item similarities of all data items constituting the content blocks. When the similarities of the data watermarks of the content blocks satisfy the first threshold range, S4 is executed. Otherwise, the embedding ratio and/or position of the data watermarks in the content blocks that do not satisfy the first threshold range is adjusted, and S2 is executed, including:

Secondary similarity calculation is performed based on the data entry similarity of all data entries constituting the content block, i.e., content block data watermark similarity. The size of the content block is set by the user according to the specific business scenario. For example, for ease of reference, the size of the content block can be set to 20 lines, 50 lines, or 100 lines.

Taking the data with a content block size of N rows as an example, the similarity of the data watermark entry is evaluated according to the method provided in S2, and the evaluation result is recorded as C. Then the secondary similarity of the content block after all the data watermarks are embedded is

Where: δ represents the content block data watermark similarity; N represents the total number of data entries in the content block; _Ci represents the data entry similarity of the i-th data entry.

Determine whether the data watermark similarity of each content block meets the first threshold range. When the data watermark similarity of each content block meets the first threshold range, execute S4. Otherwise, adjust the embedding ratio and/or position of the data watermark in the content block that does not meet the first threshold range, and execute S2.

In this embodiment, the method adopted when the similarity of the content blocks does not meet the threshold range is specifically introduced:

Method 1: Dynamically adjust the proportion of data watermark addition, including:

When the content block data watermark similarity is greater than the maximum value in the first threshold range, reducing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

When the content block data watermark similarity is less than the minimum value in the first threshold range, increasing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

The above process specifically includes: when the secondary similarity of a data content block fully embedded with a data watermark exceeds the maximum value in the first threshold range, the secondary similarity before and after the data watermark is embedded can be ensured by reducing the embedding ratio of the data watermark. For example, the embedding ratio of the data watermark can be set to 50%, 30% or 20%, etc.

When the secondary similarity of a data content block after embedding the data watermark is less than the minimum value in the first threshold range, the embedding ratio of the data watermark can be increased to increase the embedding capacity of the data watermark as much as possible. For example, the embedding ratio of the data watermark can be set to 20%, 30% or 50%, etc.

When the data items that make up the content block contain multiple field types, you can use method 2 to dynamically adjust the location where the data watermark is added, including:

Remove the original data watermark in the data entry, and embed data watermarks matching the field type into various types of fields in the data entry according to a preset ratio;

The data entries after embedding the data watermark matching the field type are evaluated for data entry similarity, the position of the field with the greatest data entry similarity is selected as the optimal position for embedding the data watermark, and the data watermark is embedded at the optimal position.

In this embodiment, the position where the data watermark is added is adjusted, specifically including: when the data entry in a data content block includes both numerical values and text and natural language, the data watermark can be added to the numerical field, or the text field, or the natural language field according to a fixed data watermark embedding ratio, and the similarity of the entries after the embedded field, or the text, or the natural language data watermark is calculated according to the method provided by S2, the position with the largest entry similarity is selected as the optimal position for adding the watermark, and the original data added in the entry is deleted, and the secondary similarity is calculated according to the entry similarity, so as to maximize the data watermark embedding capacity while ensuring that the secondary similarity after the embedded data watermark meets the threshold.

S4 calculates the similarity of the target data as a whole based on the similarity of the data watermarks of all the content blocks constituting the target data, and when the similarity of the target data as a whole satisfies the second threshold range, completes the embedding of the data watermark, otherwise adjusts the embedding ratio and/or position of the data watermark in one or more content blocks, and executes S2, including:

When the secondary similarity after embedding the data watermark meets the threshold range and the data watermark embedding capacity is increased as much as possible, the similarity of the data watermark embedded in the target data as a whole is calculated according to the secondary similarity of all content blocks, that is, the tertiary similarity. When the tertiary similarity meets the second threshold range, the embedding of the data watermark is completed, otherwise the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and S2 is executed.

When the third-level similarity does not meet the second threshold range, dynamic adjustment can be performed by adjusting the ratio as follows:

When the overall similarity of the target data is greater than the maximum value in the second threshold range, reducing the proportion of the data watermark embedded in one or more content blocks to a preset proportion;

When the overall similarity of the target data is less than the minimum value in the second threshold range, the proportion of the data watermark embedded in one or more content blocks is increased to a preset proportion.

That is, when the overall similarity of the target data does not satisfy the second threshold range, it is necessary to adjust the proportion of the data watermark embedded in one or more content blocks to a preset proportion.

When the third-level similarity does not meet the second threshold range, and in the content block to be adjusted, the data entries that constitute the content block contain multiple types of fields, the position of the data watermark embedded in the data entry can be adjusted so that the third-level similarity meets the second threshold range, thereby completing the data watermark embedding process.

This embodiment takes the example of dividing a target data into M content blocks as a whole, embedding data watermarks in each content block, and then evaluating the similarity of the data watermark entries. Then, the similarity of the content blocks is evaluated based on the similarity of the data watermark entries. The evaluation result is recorded as δ. Then, the third-level similarity of the target data content blocks after all the data watermarks are embedded is

If θ exceeds the maximum value of the set second threshold range, the embedding ratio and/or position of the data watermark is adjusted to increase the δ value and then the θ value, thereby ultimately improving the similarity of the data watermark after embedding the data as a whole.

If the minimum value of the second threshold range set by the θ distance is large, the embedding ratio and/or position of the data watermark is adjusted to increase the embedding ratio of the data watermark, so as to maximize the data watermark embedding capacity while ensuring the third-level similarity after embedding the data watermark.

In order to achieve the purpose of high concealment and high simulation after the data watermark is embedded in the target data, the embodiment of the present invention selects the appropriate proportion and distribution strategy of watermark addition according to the similarity evaluation results of different data watermark algorithms, so as to ultimately achieve high concealment and high simulation after the data watermark is embedded.

Embodiment 2: Based on the same inventive concept, the present invention also provides a data watermark embedding system for target data, as shown in FIG2 , including:

An embedding module, used for dividing the target data to be embedded with a data watermark into a plurality of content blocks, and embedding a data watermark in each content block;

A data entry similarity evaluation module is used to use a preset data similarity evaluation model to perform data entry similarity evaluation on the data entry after the data watermark is embedded;

A content block similarity evaluation module, used to evaluate the similarity of data watermarks of the content blocks based on the data entry similarities of all data entries constituting the content blocks, and when the similarities of the data watermarks of the content blocks all meet the first threshold range, the overall similarity evaluation module is executed, otherwise, the embedding ratio and/or position of the data watermarks in the content blocks that do not meet the first threshold range is adjusted, and the data entry similarity evaluation module is executed;

The overall similarity evaluation module is used to calculate the overall similarity of the target data based on the data watermark similarity of all content blocks that constitute the target data. When the overall similarity of the target data meets the second threshold range, the embedding of the data watermark is completed; otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and the data entry similarity evaluation module is executed.

On the one hand, the system uses a data similarity assessment model to evaluate the similarity of data watermark entries, data watermark embedded content blocks, and data watermark embedded data as a whole. On the other hand, based on the evaluation results, the system dynamically adjusts the proportion and distribution of watermark additions to meet the similarity threshold of the circulating data set by the user, thereby ensuring the concealment and high simulation of data watermark embedding as a whole.

In an embodiment, the system further comprises an adjustment module for adjusting the embedding ratio and/or position of the data watermark.

The adjustment module comprises:

A first adjustment unit, used for adjusting the embedding ratio of the data watermark when the data entry contains a single type of field;

The second adjustment unit is used to adjust the embedding ratio and/or position of the data watermark when the data entry contains multiple types of fields.

The adjustment module further includes: a ratio adjustment unit, which is specifically used to:

When the content block data watermark similarity is greater than the maximum value in the first threshold range, reducing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

When the content block data watermark similarity is less than the minimum value in the first threshold range, increasing the proportion of data watermarks embedded in the data entries of the content block data to a preset proportion;

When the overall similarity of the target data is greater than the maximum value in the second threshold range, reducing the proportion of the data watermark embedded in one or more content blocks to a preset proportion;

When the overall similarity of the target data is less than the minimum value in the second threshold range, the proportion of the data watermark embedded in one or more content blocks is increased to a preset proportion.

The adjustment module further includes: a position adjustment unit, specifically used for:

Remove the original data watermark in the data entry, and embed data watermarks matching the field type into various types of fields in the data entry according to a preset ratio;

The data entries after embedding the data watermark matching the field type are evaluated for data entry similarity, the position of the field with the greatest data entry similarity is selected as the optimal position for embedding the data watermark, and the data watermark is embedded at the optimal position.

In an embodiment, the field types in the data entry include any one or more of the following:

Numeric fields, text fields, and natural language fields.

In an embodiment, the embedded module is specifically used for:

When the data entry includes a numerical field, embedding a numerical data watermark in the numerical field;

When the data entry includes a text field, embedding a data watermark of a character text type in the text field;

When the data entry includes a natural language field, a natural language data watermark is embedded in the natural language field.

In the embodiment, the data entry similarity assessment module is specifically used to:

When a numerical data watermark is embedded in the numerical field of the data entry, the numerical values before and after the data watermark is embedded are deconstructed and segmented, and the data entry similarity is evaluated using a Euclidean distance vector data similarity evaluation model;

When a character text type data watermark is embedded in the text field of the data entry, the ASCII code values before and after the data watermark is embedded are deconstructed, and the data entry similarity is evaluated by a cosine vector data similarity evaluation model;

When a natural language data watermark is embedded in the natural language field of the data entry, a space vector model is applied to the natural language field before and after the data watermark is embedded to perform word deconstruction and segmentation, and the deconstruction and segmentation results are evaluated for data entry similarity using a cosine vector data similarity evaluation model.

In the embodiment, the content block data watermark similarity evaluation is performed according to the following formula:

Where: δ represents the content block data watermark similarity; N represents the total number of data entries in the content block; _Ci represents the data entry similarity of the i-th data entry.

In the embodiment, the similarity of the target data as a whole is evaluated as follows:

Where: θ represents the similarity of the target data as a whole; M represents the total number of content blocks in the target data; _δi represents the watermark similarity of the content block data of the i-th content block.

Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment in combination with software and hardware. Moreover, the present application may adopt the form of a computer program product implemented in one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) that include computer-usable program code.

The present application is described with reference to the flowchart and/or block diagram of the method, device (system) and computer program product according to the embodiment of the present application. It should be understood that each process and/or box in the flowchart and/or block diagram, and the combination of the process and/or box in the flowchart and/or block diagram can be realized by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device produce a device for realizing the function specified in one process or multiple processes in the flowchart and/or one box or multiple boxes in the block diagram.

These computer program instructions may also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory produce a manufactured product including an instruction device that implements the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

These computer program instructions may also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, whereby the instructions executed on the computer or other programmable device provide steps for implementing the functions specified in one or more processes in the flowchart and/or one or more boxes in the block diagram.

The above are merely embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention are included in the scope of the claims of the present invention to be approved.

Claims (11)

1. A data watermark embedding method of target data, characterized in that, comprising: S1 divides the target data to be embedded with the data watermark into multiple content blocks, and embeds the data watermark in the data entry of each content block; S2 uses a pre-set data similarity evaluation model to evaluate the similarity of data items on the data items before embedding the data watermark and the data items after embedding the data watermark; S3 evaluates the similarity of content block data watermarks based on the data entries of all data items that make up the content block before embedding the data watermark and the data entries of all the data items that make up the content block after embedding the data watermark. When the data watermarks of each content block are similar When all degrees meet the first threshold range, execute S4, otherwise adjust the embedding ratio and/or position of the data watermark in the content block that does not meet the first threshold range, and execute S2; S4 is based on the data integrity of all the content blocks that make up the target data before embedding the data watermark, and the data integrity of all the content blocks that make up the target data after embedding the data watermark, perform the similarity calculation of the target data overall, when the target data overall When the similarity of the content meets the second threshold range, the embedding of the data watermark is completed; otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and S2 is executed. 2. The method according to claim 1, wherein adjusting the embedding ratio and/or position of the data watermark comprises: When a data entry contains a single type of field, adjust the embedding ratio of the data watermark; When the data entry contains multiple types of fields, adjust the embedding ratio and/or position of the data watermark. 3. The method according to claim 2, wherein said adjusting the embedding ratio of the data watermark comprises: When the content block data watermark similarity>the maximum value in the first threshold range, then reduce the ratio of the data watermark embedded in the data entry of the content block data to a preset ratio; When the content block data watermark similarity is less than the minimum value in the first threshold range, then increase the ratio of the data watermark embedded in the data entry of the content block data to a preset ratio; When the overall similarity of the target data>the maximum value in the second threshold range, then reduce the proportion of the data watermark embedded in one or more content blocks to a preset proportion; When the overall similarity of the target data is less than the minimum value in the second threshold range, then increase the proportion of the data watermark embedded in one or more content blocks to a preset proportion. 4. The method according to claim 2, wherein said adjusting the embedding position of the data watermark comprises: Remove the original data watermark in the data entry, and embed the data watermark matching the field type into various types of fields in the data entry according to the preset ratio; Evaluate the similarity of the data entry after embedding the data watermark that matches the field type, select the location of the field with the largest similarity of the data entry as the optimal location for embedding the data watermark, and embed the data watermark at the optimal location. 5. The method according to any one of claims 2 or 4, wherein the field types in the data entry include any one or more of the following: Numeric fields, text fields, and natural language fields. 6. The method according to claim 5, wherein embedding a data watermark in the data entry comprises: When the data entry includes a numerical field, embedding a numerical data watermark in the numerical field; When the data entry includes a text field, embedding a character text data watermark in the text field; When the data entry includes a natural language field, a natural language data watermark is embedded in the natural language field. 7. The method according to claim 1, wherein said adopting a preset data similarity evaluation model to evaluate the data entry similarity of the data entry after the embedded data watermark comprises: When a numerical data watermark is embedded in the numerical field of the data entry, the numerical value before and after the data watermark is embedded is deconstructed and word-segmented, and the data entry similarity is evaluated through the Euclidean distance vector data similarity evaluation model; When the text field of the data entry is embedded with a character text type data watermark, the ASCII code value before and after the data watermark is embedded is deconstructed, and the similarity evaluation of the data entry is performed through the cosine vector data similarity evaluation model; When a natural language-type data watermark is embedded in the natural language field of the data entry, the space vector model is applied to the natural language field before and after the data watermark is embedded to perform deconstruction and segmentation, and the deconstruction and segmentation results are evaluated by the cosine vector data similarity model Perform data entry similarity evaluation. 8. The method according to claim 1, wherein the content block data watermark similarity evaluation is carried out according to the formula:

In the formula: δ represents the similarity of the content block data watermark; N represents the total number of data entries in the content block; C _i represents the data entry similarity of the i-th data entry. 9. The method according to claim 1, wherein the overall similarity of the target data is evaluated as follows:

In the formula: θ represents the overall similarity of the target data; M represents the total number of content blocks in the target data; δ _i represents the similarity of the content block data watermark of the i-th content block. 10. A data watermark embedding system for target data, comprising: An embedding module, configured to divide the target data to be embedded with the data watermark into multiple content blocks, and embed the data watermark in the data entry of each content block; The data entry similarity evaluation module is used to evaluate the data entry similarity between the data entry before embedding the data watermark and the data entry after embedding the data watermark; The content block similarity evaluation module is used to perform content block data watermark similarity on the data items of all data items that make up the content block before the data watermark is embedded, and the data items of all the data items that make up the content block after the data watermark is embedded Evaluate, when the similarity of each content block data watermark meets the first threshold range, execute the overall similarity evaluation module, otherwise adjust the embedding ratio and/or position of the data watermark in the content block that does not meet the first threshold range, and execute the described Data entry similarity evaluation module; The overall similarity evaluation module is used to evaluate the overall similarity of the data integrity based on all content blocks that make up the target data before embedding the data watermark, and the data integrity of all content blocks that make up the target data after embedding the data watermark. When the overall similarity of the data meets the second threshold range, the embedding of the data watermark is completed; otherwise, the embedding ratio and/or position of the data watermark in one or more content blocks is adjusted, and the data item similarity evaluation module is executed. 11. The system according to claim 10, wherein the data entry similarity evaluation module is specifically used for: When a numerical data watermark is embedded in the numerical field of the data entry, the numerical value before and after the data watermark is embedded is deconstructed and word-segmented, and the data entry similarity is evaluated through the Euclidean distance vector data similarity evaluation model; When the text field of the data entry is embedded with a character text type data watermark, the ASCII code value before and after the data watermark is embedded is deconstructed, and the similarity evaluation of the data entry is performed through the cosine vector data similarity evaluation model; When a natural language-type data watermark is embedded in the natural language field of the data entry, the space vector model is applied to the natural language field before and after the data watermark is embedded to perform deconstruction and segmentation, and the deconstruction and segmentation results are evaluated by the cosine vector data similarity model Perform data entry similarity evaluation.

Images