Vector Font Watermarking via Line Width Ratio Adjustment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for embedding watermarks in vector fonts are not robust or transparent, as they either alter the font quality significantly or are removed by commercial font tools, and existing techniques are not suitable for vector fonts due to scalability issues.
Innovation Solution
A method that adjusts the width ratio of selected line pairs within specific width ranges to embed binary watermark information, ensuring the changes are imperceptible and robust against scaling, using a secret key for protection and error detection mechanisms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If additional points are inserted into vector font lines to embed watermark, then watermark information can be embedded, but the inserted points are removed by commercially available font tools
Solution Approach 1:
The patent changes the parameters of existing vector font elements (line widths, character spacing, point positions) rather than inserting new elements. By modifying these parameters within acceptable tolerances, the watermark is embedded in a way that survives font tool processing while maintaining visual quality.
Solution Approach 2:
The patent applies different modification strategies to different parts of the vector font. Critical lines that define character shape are modified minimally, while less critical lines can tolerate greater modifications for watermark embedding. This localized approach preserves overall font quality while embedding watermark information.
2Loss of information
If individual vector points are shifted to introduce watermark bits, then watermark information can be embedded, but the changes seriously affect the quality of the material
Solution Approach 1:
Instead of shifting individual points which causes visible distortion, the patent modifies line width parameters and character spacing parameters. These parameter changes are subtler and can be applied within tolerances that maintain visual quality while still embedding detectable watermark information.
Solution Approach 2:
The patent applies minimal necessary changes to vector parameters - just enough to embed the watermark bits but not so much as to degrade visual quality. By using partial action rather than excessive point shifting, the watermark remains detectable while the font quality is preserved.
3Loss of information
If spacing between characters or words is changed to introduce watermark, then watermark information can be embedded, but font scaling has strong impact on the changed spacing affecting legibility
Solution Approach 1:
The patent creates asymmetric modifications within symmetric structures. By selectively modifying spacing or line widths in specific patterns that break symmetry, the watermark information is encoded in a way that is invariant to uniform scaling operations, making the watermark robust against font scaling while still detectable.
4Reliability
If line width ratios are changed to embed watermark information, then binary watermark information can be embedded robustly, but the changes must remain imperceptible
Solution Approach 1:
The patent modifies line width parameters within acceptable tolerances. By changing line width ratios subtly rather than dramatically, the watermark information is embedded in a robust manner while the modifications remain imperceptible to human observers and do not degrade font quality.
Data Source
Figure 1~2
Figure 3~4
AI summary
The method involves randomly or pseudo-randomly selecting two longitudinal lines (10, 12) of a vector character or a line of two vector characters with breadths belonging to a common group of line breadths, and defining the two lines. Breadth relationship of the two lines is changed in such a manner that one of the lines is broader and narrower than the other line, where two binary water-mark information are embedded into the vector characters.