Abstract | ||
---|---|---|
Traditional digital image watermarking schemes have the inherent conflicts between imperceptibility and robustness because the insertion of watermarks into the host image inevitably introduces some perceptible quality degradation. Image zero-watermark techniques resolve this dilemma by extracting invariant features from its host image. At present, most zero-watermark schemes are more robust to geometric attacks and the common signal process. But they are not robust to print-cam attack, where the image is first printed and then captured with a mobile phone. To solve the problem, this paper addresses a novel zero-watermarking scheme based on improved ORB (Oriented FAST and Rotated BRIEF) features. Firstly, we increase the initial ORB matching points to improve the matching rate and add color information as another feature. So the added color features can remove the false-positive matches which the color correspondence is wrong. Then, ORB keypoints, ORB descriptors and color descriptors are the zero-watermark to be registered. Finally, we perform outlier filtration using the RANSAC (random sample consensus) method and obtain the final inliers during the zero-watermark detection. Then the original image match the one captured with a mobile phone by the zero-watermark. Experimental results show that the proposed image zero-watermark scheme is robust to geometric attacks, signal processing attacks and print-cam attack. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1007/978-3-319-19321-2_14 | DIGITAL-FORENSICS AND WATERMARKING, IWDW 2014 |
Keywords | Field | DocType |
Zero-watermark, Feature match, ORB, Color features, Random sample consensus | Signal processing,Computer vision,Digital watermarking,Computer science,RANSAC,Orb (optics),Outlier,Robustness (computer science),Artificial intelligence,Invariant (mathematics),Mobile phone | Conference |
Volume | ISSN | Citations |
9023 | 0302-9743 | 0 |
PageRank | References | Authors |
0.34 | 9 | 6 |
Name | Order | Citations | PageRank |
---|---|---|---|
Jianfeng Lu | 1 | 24 | 3.78 |
Qianru Huang | 2 | 5 | 0.78 |
Meng Wang | 3 | 0 | 0.34 |
Li Li | 4 | 237 | 34.83 |
Junping Dai | 5 | 5 | 0.78 |
Chin Chen Chang | 6 | 7849 | 725.95 |