2000 Fiscal Year Final Research Report Summary
Research on Digital Watermarking for Video Data and Text Documents
| Project/Area Number |
11650386
|
|---|
| Research Category |
Grant-in-Aid for Scientific Research (C)
|
| Allocation Type | Single-year Grants |
|---|
| Section | 一般 |
|---|
| Research Field |
情報通信工学
|
|---|
| Research Institution | KYUSHU UNIVERSITY |
|---|
Principal Investigator |
MIYAZAKI Akio Graduate School of Information Science and Electrical Engineering, KYUSHU UNIVERSITY Ass.Prof., 大学院・システム情報科学研究院, 助教授 (70192763)
|
|---|
| Project Period (FY) |
1999 – 2000
|
| Keywords | digital watermark / information security / digital contents / digital video / text document / copyright protection / wavelet transform / morphological signal processing |
|---|
| Research Abstract |
(1) Development of digital watermarking method for images and video :
We have proposed a watermarking method for both still images and video using the wavelet transform. First, we have applied the proposed method to the watermark for still images and have shown that this method is robust against lossy compression like JPEG, adding noise, reduction of grayscale level, and StirMark attacks. Next, we have applied this method to the watermark for video. As a result, we have confirmed that the watermark is extracted from the watermarked video compressed with MPEG.
(2) Development of digital watermarking method for text documents :
We have presented a watermarking method for text documents using the morphological signal processing technique. This method embeds information data (watermark) into binary patterns such as character and/or figure patterns in text document. We have applied this method to the letters of the English alphabet and realize a watermark technique for English documents. Using the proposed technique, we have confirmed from numerical experiments that the embedded data can be detected from watermarked letters with noise and printed on paper.
(3) On the evaluation of performance of digital watermarking method in the frequency domain :
We have focused on the watermarking method in the frequency domain, in which watermark is embedded using a controlled quantization process, and suggested the model of watermark embedding and detecting system. Next, using this model, we have analyzed the watermarking method and proposed how to evaluate its performance. As a result, we can know how many bits are able to be embedded and how large the embedded intensity should be set. Furthermore, we have discussed the difference of the performance between the DWT- and DCT-based watermarking methods.
|
Research Products
(17 results)
