ANiTW: A Novel Intelligent Text Watermarking technique for forensic identification of spurious information on social media

Digital Watermarking is required in multimedia applications where access to sensitive information has to be protected against malicious attacks. Since the digital text is one of the most widely used digital media on the Internet, the significant part of Web sites, social media, articles, eBooks, and so on is only plain text. Thus, copyrights protection of plain-texts is still a remaining issue that must be improved to provide proof of ownership and verify content integrity of vulnerable digital texts. In this research, we propose a novel intelligent text watermarking technique called ANiTW which utilizes an instance-based learning algorithm to hide an invisible watermark into Latin text-based information such that the hidden watermark can be extracted, even if a malicious user manipulates a portion of the watermarked information. Extensive experiments demonstrate the superior efficiency of the ANiTW with a significant improvement especially in the short text domain. To the best of our knowledge, this is the first intelligent text watermarking technique that provides an invisible signature for forensic identification of spurious information on social media by evaluating the manipulation rate of watermarked information, while the other existing approaches only consider the robust/fragile marking of signature into cover text.