Semantic-driven watermarking of relational textual databases

• Published in: Expert systems with applications 2021-04, Vol.167, p.114013, Article 114013
• Main Authors: Pérez Gort, Maikel Lázaro, Olliaro, Martina, Cortesi, Agostino, Feregrino Uribe, Claudia
• Format: Article
• Language: English
• Subjects: Digital watermarking; Perturbation; Relational data bases; Relational databases; Semantic similarity analysis; Semantics; Substitutes; Watermarking
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2020.114013

•Mark embedding through synonyms substitutions avoids to compromise data quality.•Controlling the number of times each mark is embedded contributes to high resiliency.•Multi-word textual attributes in relational data helps to protect their ownership.•Majority voting compensates lack of precision in word sense disambiguation algorithms. In relational database watermarking, the semantic consistency between the original database and the distorted one is a challenging issue which is disregarded by most watermarking proposals, due to the well-known assumption for which a small amount of errors in the watermarked database is tolerable. We propose a semantic-driven watermarking approach of relational textual databases, which marks multi-word textual attributes, exploiting the synonym substitution technique for text watermarking together with notions in semantic similarity analysis, and dealing with the semantic perturbations provoked by the watermark embedding. We show the effectiveness of our approach through an experimental evaluation, highlighting the resulting capacity, robustness and imperceptibility watermarking requirements. We also prove the resilience of our approach with respect to the random synonym substitution attack.