Text Watermarking via Typographical Perturbation for Document Tracking
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The sharing of electronic files poses a threat to data integrity due to unauthorized disclosures, making it difficult to identify leaked documents and the entities responsible for their unauthorized dissemination, as metadata can be stripped or modified, and typographical properties are not preserved in printed, scanned, or screen-captured documents.
Innovation Solution
A computer-implemented method for encoding and decoding digital watermarks by formatting text elements in digital documents with multi-bit patterns derived from identifiers, where the identifiers are embedded as perturbations in typographical properties like height and width, allowing for accurate retrieval even from scanned documents by measuring these properties and using delimiters to select appropriate decoding schemes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If metadata is used to identify documents, then document identification is simplified, but metadata can be stripped or modified making identification unreliable
Solution Approach 1:
The identifier is segmented and embedded across multiple text elements rather than stored as a single metadata field. Each text element contains a portion of the identifier encoded in its typographical properties, distributing the identification information throughout the document structure to prevent complete loss if any single element is modified.
Solution Approach 2:
Instead of relying on metadata that can be stripped, the identifier information is copied into the typographical properties of text elements themselves. This creates redundant copies of the identification data within the document's visual structure, ensuring persistence even when metadata is removed.
2Reliability
If typographical properties are modified to encode identifiers, then document tracking capability is enhanced, but document fidelity and readability may be compromised
Solution Approach 1:
Only specific text elements are modified with encoded identifiers while leaving other portions of the document unchanged. The encoding is applied locally to selected words or phrases, preserving the overall document integrity and readability while embedding tracking information in discrete locations.
Solution Approach 2:
The identifier is encoded by making subtle changes to typographical parameters such as font size, weight, or spacing rather than altering the visible text content. These parameter changes are imperceptible to human readers but detectable by decoding systems, maintaining both document fidelity and tracking capability.
3Loss of information
If multiple text elements are modified to carry identifier portions, then encoding capacity increases, but decoding complexity increases
Solution Approach 1:
The decoding process uses feedback mechanisms where the system reads text elements sequentially, accumulates identifier portions, and validates the reconstructed identifier against expected formats. This feedback loop enables error detection and correction, ensuring complete identifier retrieval even when some text elements are modified or missing.
Solution Approach 2:
The encoding scheme incorporates preliminary organization of identifier segments across text elements with predefined positions and patterns. This preliminary structuring simplifies the decoding process by providing a predictable sequence for reassembling the identifier, reducing computational complexity despite multiple modified elements.
Data Source
AI summary
Described are methods and systems that watermark text files in document or image formats using efficient encoding schemes. A unique identifier is encoded into a document by perturbing typographical properties of document elements, such as the lengths and widths of words, lines, or spaces, to encode multiple bits per element. Perturbations to the rendered dimensions of elements create patterns, digital watermarks, that can be decoded to recover the unique identifier, which can in turn be used to identify a user who disclosed or was otherwise responsible for a leaked document.