2004 Fiscal Year Final Research Report Summary
An Improvemet of Watermarking Methods for Images and Video using Deconvolution Techniques
| Project/Area Number |
15560343
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
Communication/Network engineering
|
|---|
| Research Institution | Kyushu Sangyo University |
|---|
Principal Investigator |
MIYAZAKI Akio Kyushu Sangyo University, Faculty of Information Science, Professor, 情報科学部, 教授 (70192763)
|
|---|
| Project Period (FY) |
2003 – 2004
|
| Keywords | digital watermark / watermarking / information security / image / video / evaluation of watermark / convolution / deconvolution |
|---|
| Research Abstract |
1.On the evaluation of wavelet filter banks for wavelet-based image watermarking : In the wavelet-based watermarking, similar to the wavelet-based coding, the choice of wavelet filter banks generally affects the performance of watermarking methods. Hence a natural question arises for the wavelet-based watermarking. What is the best wavelet filter bank for use ? In this research, we have evaluated wavelet fitter banks used in watermarking theoretically from the viewpoint of the quality of watermarked image and the robustness of watermark, and then tried to answer the best wavelet filter bank problem in the wavelet-based watermarking. Based on the results, we have discussed the performance of the watermarking method using orthogonal wavelet and the watermarking method using biorthogonal wavelet, and carried out the comparison between the two methods.
2.An improvement of the watermark detection in the correlation-based image watermarking methods : This research aims to improve the accuracy
… More
of watermark detection in correlation-based watermarking methods for images. In the correlation-based watermarking method, watermark detection is carried out by computing the correlation between a watermarked image and a pseudorandom code to be checked for the presence. Hence its accuracy depends on the characteristics of host images. In this research, we have firstly shown an experimental result that a certain image may cause watermark detection errors. Then we have developed an improved version of the watermark embedding process to be effective in such a situation. The improvement is accomplished by designing a whitening filter using a nonlinear programming algorithm. Numerical experiments have shown good performance as expected by us.
3.An improvement of watermark extracting process using the deconvolution technique: In this research, we have focused on digital watermarking techniques based on modifying the spectral coefficient of an image. We have presented a model of the watermark embedding and extraction processes and described a model of watermark distortion caused by image processing and attack of watermarked image as a convolution model. Based on the model, we have clarified the reason why detection errors occur in the watermark extraction process and formulated the problem of the improvement of the watermark detection as a kind of blind deconvolution problem. Then we have studied the watermark detection problem using Bayesian estimation technique, and as a solution to this problem, proposed a watermark detection method based on the EM (Expectation-Maximization) algorithm. Less
|
