A Semi-Blind Robust DCT Watermarking Approach for Sensitive Text Images

作者：Laouamer Lamri; Tayan Omar^*

来源：Arabian Journal for Science and Engineering, 2015, 40(4): 1097-1109.

DOI：10.1007/s13369-015-1596-y

• Semi-blind watermarking
• Robustness
• DCT
• Quran text image
• Authentication
• Attacks

This paper presents a new approach to address the challenges associated with integrity verification, proof of authenticity, tamper detection and assurance of copyright protection for digital images propagated online. In particular, the problem addressed here primarily relates to the protection of sensitive online textual images that may be targeted for counterfeiting or intentional/unintentional modifications which would result with a fake copy being disseminated. Typically, sensitive text images are characteristic of further challenges and constraints on embedding/extraction techniques compared to their more general digital picture/image counterparts. With many digital protection schemes addressing the problem of less-sensitive content comprising of a rich set of colours and textures, this paper proposes a novel watermarking approach for the case of sensitive text images in order to confirm content originality and integrity verification under some known constraints of the image type. Hence, a new approach based on discrete cosine transform and linear interpolation is proposed and applied on digital Quran text images, which are considered as ideal samples of content with stringent sensitivity requirements. Detailed results had presented the significance of this approach and had demonstrated the high robustness obtained when faced with various attack scenarios that include rotational attacks, JPEG compression attacks, noising attacks and median filtering attacks. Significantly, a major contribution of this paper was found in the novel watermark extraction approach which had enabled perfect watermark extraction under a range of attack scenarios. Finally, a significant advantage of the proposed approach was found in its broad applicability for authentication and protection of other sensitive and less-sensitive digital image content.

• 出版日期2015-4