Palindromic Fingerprint Embedding for Robust File Authorship Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current steganographic techniques for authorship protection and copy detection of digital files lack resilience to modifications and transformations, making them vulnerable to plagiarism.
Innovation Solution
Embedding palindromic representations of digital fingerprints in the least significant bits of binary data, using orthogonal interspersing and asymmetric delimiters, to enhance detectability and diagnose alterations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional steganographic techniques are used to embed digital fingerprints, then the embedding process is simple, but the detectability of the fingerprint decreases when the file is transformed
Solution Approach 1:
The patent embeds palindromic sequences (which are symmetric) within asymmetric delimiters. The palindrome structure allows the fingerprint to be detected regardless of the direction of scanning, making it resilient to transformations like flipping or rotating the file. The asymmetric delimiters provide orientation information to determine the original orientation of the file.
Solution Approach 2:
The digital fingerprint is divided into multiple palindromic sequences that are distributed throughout the file. Each palindrome can be independently detected, and the collection of palindromes collectively identifies the file's authorship. This segmentation allows the fingerprint to remain detectable even if some portions of the file are transformed or modified.
2Reliability
If the least significant bits are replaced with fingerprint data, then the steganographic embedding is achieved, but the resilience to modifications and transformations is limited
Solution Approach 1:
The patent performs preliminary actions by first identifying suitable least significant bits for replacement, then embedding the palindromic sequences in a structured manner. The delimiters are placed beforehand to mark the positions and orientations of the palindromes. This preliminary structuring ensures that the embedded fingerprint remains resilient to modifications while maintaining a manageable embedding process.
3Reliability
If palindromic repeats are embedded in the data file, then the detectability increases under transformations, but the file size increases
Solution Approach 1:
The patent embeds palindromic sequences only in specific locations within the file, namely in the least significant bits of selected data elements. The delimiters are strategically placed to mark only the necessary positions. This localized embedding approach increases detectability while minimizing the increase in file size, as the palindromes are embedded sparsely rather than throughout the entire file.
Data Source
AI summary
The invention is notably directed to a computer-implemented method of processing a data file, wherein the method comprises: accessing (S10) a data file (F1) containing binary data that represent analogue data, to populate (S20) a first data structure (11) with the binary data; obtaining (S30) a second data structure (12) by replacing, in the first data structure (11), data (111) corresponding to some of least significant bits of the binary data with substitutive bit values (121); obtaining (S40) a digital fingerprint of the second data structure (12); forming (S50) a binary representation of a payload that includes repeats of a sequence, the sequence containing a palindrome based on the digital fingerprint; obtaining (S60) a third data structure (13) by embedding the binary representation of the payload (131, 132) formed in the second data structure (12), in place of the substitutive bit values (121); and storing (S70) a modified data file (F1m) including binary data of the third data structure (13). The invention is further directed to related detection methods, data files, computerized systems, and computer program products.