Digital Watermarking for Textual Data Tables
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing digital watermarking techniques for structured textual data are vulnerable to subset attacks, primary key deletion/alteration attacks, and significantly alter the data content, making them unsuitable for statistical and machine learning analysis.
Innovation Solution
A method for embedding digital watermarks in structured textual data by selecting a subset of cells, determining a primary cell key and cell partition number, and embedding a digital watermark ID code based on these, using a secret key common to all cells, which allows for negligible data modification and blind extraction without relying on primary keys.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing structured textual data watermarking solutions are applied, then copyright protection and ownership proof are achieved, but the data content is substantially altered making it unsuitable for statistical and machine learning analysis
Solution Approach 1:
The patent applies local quality by selectively embedding watermarks only in certain cells rather than uniformly across all data. The system determines a 'watermark strength' for each cell based on its sensitivity to modification, allowing high-protection cells to be watermarked while preserving integrity in analysis-critical cells. This resolves the contradiction by making protection localized rather than global.
Solution Approach 2:
The patent implements partial action by embedding watermarks in only a subset of cells rather than all cells. The system uses a sensitivity analysis to identify which cells can tolerate watermarking without compromising data usability for analytics. This partial application of watermarking maintains copyright protection while preserving data integrity for machine learning and statistical analysis.
2Reliability
If cell-level watermarking is applied to structured textual data, then subset attacks are prevented, but the data modification is significant negatively impacting analytical usefulness
Solution Approach 1:
The patent applies local quality by selectively embedding watermarks only in certain cells rather than uniformly across all data. The system determines a 'watermark strength' for each cell based on its sensitivity to modification, allowing high-protection cells to be watermarked while preserving integrity in analysis-critical cells. This resolves the contradiction by making protection localized rather than global.
Solution Approach 2:
The patent implements partial action by embedding watermarks in only a subset of cells rather than all cells. The system uses a sensitivity analysis to identify which cells can tolerate watermarking without compromising data usability for analytics. This partial application of watermarking maintains copyright protection while preserving data integrity for machine learning and statistical analysis.
3Ease of operation
If primary key attribute is used for partitioning in watermarking, then watermark embedding and extraction is simplified, but the system becomes vulnerable to deletion and alteration attacks
Solution Approach 1:
The patent applies segmentation by dividing the data into sensitivity groups rather than relying on a single primary key. Each group is assigned a sensitivity level that determines watermark embedding behavior. This segmentation approach maintains operational simplicity while improving security by not depending on a single vulnerable primary key attribute that can be deleted or altered.
Solution Approach 2:
The patent changes the parameter used for partitioning from primary key attribute to data sensitivity level. Instead of using the primary key which can be deleted or modified, the system uses a sensitivity parameter that is determined through analysis of each cell's importance to data integrity and analytical usefulness. This parameter change makes the system resistant to deletion and alteration attacks while maintaining ease of operation.
Data Source
AI summary
Methods and system for embedding digital watermark information into textual data arranged in a table of cells are provided. A first subset of cells are selected and for each primary cell key and cell partition number are determined. A portion of a digital watermark ID code is embedded at an embedding position determined based on the partition number. Methods and systems for extracting digital watermark information from the textual data are also provided. A cell is fetched from the table and the presence of portion of the digital watermark ID code is determined. A primary cell key and cell partition number are determined. A portion of the digital watermark ID code is extracted at the embedding position within the cell, the embedding position determined based on the cell partition number. The digital watermarking systems and methods provide tracking for unauthorized copying of the data while modifying only a subset of the data.