Robust Digital Text Watermarking Algorithm based on Unicode Extended Characters

摘要

Objectives: A new text watermarking algorithm is proposed in this paper to protect documents from malicious attacks.Methods: In this paper, a novel digital text watermarking algorithm is developed based on Unicode extended characters. The algorithm is implemented to encompass watermark generation, embedding and extraction components. Predefined encoding tables are constructed to achieve the embedding while scrambling mechanism is proposed in generation and extraction to secure the watermarks. Findings: To evaluate the algorithm, it was tested, using ten different text samples, under various attacks such as conversion, reformatting, copying, insertions and deletions. The proposed algorithm has attained high level of imperceptibility with PSNR between 63.15 and 70.88 and SIM between 99.93% and 99.97%.Evaluating the robustness of the algorithm proved that it resists most of the attacks with high detection accuracy reached 100% in most cases. It has also achieved improved capacity with 2 bits/word and increased security level compared to the previous works. Application/Improvements: The proposed algorithm has added a new value to the information security field and can be used for documents protection from various malicious aspects.