Embedded Files
Invisibility of the Proposed Algorithm:
Original and Extracted Watermarks:
Detailed Results:
Correlation Coefficient For All Experimental Images After Attacks:
Intended Application of the Proposed Scheme:
Generally, if the watermarking literature is reviewed then one can find many applications where watermarking plays a remarkable role. These applications are either related to host media or watermark. If the applications are related to host media then the attacks are considered to be not much destructive. In other words, the host image is not destroyed after these attacks. These applications include copyright protection, proof of ownership, authentication etc. On the other hand, if the applications are related to watermark then the main concern is to protect watermark even after the extreme attacks. The best application for this scenario is covert communication (as also said by the reviewer). Covert communication can be viewed as the communication between two parties such that they are intended to hide the information. One of the possible examples comes from defence. Assume that meaningful information (in the form of text, image or video) is transmitted from one army station to another. Here, there is a need of a watermarking scheme which can protect the information even after extreme intentional and non-intentional attacks.
The proposed watermarking scheme is applicable to convert communication application along with all traditional application like copyright protection, proof of ownership, authentication etc. The simulation results appeared in manuscript have checked proposed watermarking scheme with extreme attacks (like 100:1 JPEG compression, 100% addition of noise etc.) which practically destroyed the watermarked image. For example, 100% additive noise or 100:1 JPEG compression ratio makes the watermarked images unusable but the extracted watermarks are well recognizable. Therefore, the requirement for covert communication fulfills by the simulation results presents in the manuscript.
For other traditional applications where usual attacks parameters are used instead of extreme parameters, proposed scheme again works fine. Because, if a scheme will work properly after extreme attacks then it will also work properly after usual attacks. To prove our claim, the results with usual attacks (like 30:1 JPEG compression, addition of noise with reasonable SNR etc.) are presented here. The extreme attacks which have been used in the analysis are Average & median filtering, noise addition, JPEG compression, rotation, resizing and cropping. Here, we have provided the new simulation results for these attacks with usual parameters i.e. Average & median filtering with window 5x5; Gaussian and Salt & pepper noise addition with 30% noise addition; JPEG Compression with CR 30:1; Rotation with 10 degree; Resizing with 512→128→512 and Cropping with 25% area cropped. Finally, the corresponding correlation coefficient values are given in table 1.
Results for Attacks with Usual Parameters (Here results are given for Fruits and Lady Images only)
Note: Kindly download all files to your local hard disk. Sometimes google docs may degrades the quality of the images included in the results.
Page updated
Google Sites
Report abuse