A Spatial Robust Watermarking Techniques Based on Image Statistics
碩士 === 中原大學 === 電機工程研究所 === 92 === Abstract In this thesis, we propose four spatial robust watermarking techniques based on image statistics. In this research, we use image statistics concept, adaptive block division and human visual system (HVS) to achieve a very robustness with imperceptibly and...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | en_US |
Published: |
2004
|
Online Access: | http://ndltd.ncl.edu.tw/handle/72863632759462652394 |
id |
ndltd-TW-092CYCU5442024 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-092CYCU54420242016-01-04T04:08:53Z http://ndltd.ncl.edu.tw/handle/72863632759462652394 A Spatial Robust Watermarking Techniques Based on Image Statistics 以影像統計資訊為基礎之強健性空間浮水印技術 Ming-Hui Ni 倪明輝 碩士 中原大學 電機工程研究所 92 Abstract In this thesis, we propose four spatial robust watermarking techniques based on image statistics. In this research, we use image statistics concept, adaptive block division and human visual system (HVS) to achieve a very robustness with imperceptibly and simple watermarking system. The first and second schemes are with the original image. In the first scheme, we utilize image statistics concept to obtain an embedded threshold to make adaptive modification for the whole block pixels. The main purpose of this idea is to get a robust property for spatial watermarking. The only difference of these two schemes is that the second scheme adds human visual system (HVS) and adaptive block division. The main purpose of adding human visual system is to increase the imperceptibility of watermarked image. On the other hand, the main purpose of adaptive block division is to utilize statistics concept to obtain a threshold from three blocks other than the block embedded the watermarking. Next, in the extracted algorithm, the watermark could be extracted by statistics concept with the corrupted embedded threshold which could resist the plenty attacks. It reaches the robustness of our design goal. The third and fourth schemes are without the original image. They are proposed to improve the first and second schemes. The third scheme combines the first and second schemes. The main idea to match the purpose without original image is that the threshold is embedded into three blocks other than the block embedded the watermarking. On the other hand, the fourth and second schemes use the similar concept. The main difference is that the third and fourth schemes would sacrifice a little the imperceptibility of watermarked image and the robustness of watermark to match the purpose without the original image. The contributions of our research are as follows: (1) Derive a high-speed watermarking system. (2) Reach a very well robustness property. (3) Extract watermarking without original image. They are described as follows: In recent years, the proposed schemes by most of the scholars were based on frequency-domain. The main reason is that it is easier to reach imperceptibility, robustness and without the original image than on spatial-domain, simultaneously. But it has a disadvantage that the operation speed of the whole system is slower than on spatial-domain. Therefore, we propose spatial-domain watermarking techniques to derive a high-speed watermarking system with imperceptibility, robustness and without original image. In the past, for spatial-domain, the proposed schemes by most of the scholars were with the original image. Our first and second schemes are with the original image, too. However, our proposed schemes are more robustness than traditional watermarking schemes. In addition, so far, only few of scholars proposed schemes without original image based on Spatial-Domain. However, the robustness of their proposals was poor. In this thesis, the third and fourth proposals without original image by the authors have very well robustness under some standard attacks. The purpose of without original image is that the copyright owner cannot spend the extra disk space to preserve the original image. From this study, we believe that the research results will be probably practical and efficient on applications of copyright protection. Shih-Hsiung Twu 涂世雄 2004 學位論文 ; thesis 77 en_US |
collection |
NDLTD |
language |
en_US |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 中原大學 === 電機工程研究所 === 92 === Abstract
In this thesis, we propose four spatial robust watermarking techniques based on image statistics. In this research, we use image statistics concept, adaptive block division and human visual system (HVS) to achieve a very robustness with imperceptibly and simple watermarking system.
The first and second schemes are with the original image. In the first scheme, we utilize image statistics concept to obtain an embedded threshold to make adaptive modification for the whole block pixels. The main purpose of this idea is to get a robust property for spatial watermarking. The only difference of these two schemes is that the second scheme adds human visual system (HVS) and adaptive block division. The main purpose of adding human visual system is to increase the imperceptibility of watermarked image. On the other hand, the main purpose of adaptive block division is to utilize statistics concept to obtain a threshold from three blocks other than the block embedded the watermarking. Next, in the extracted algorithm, the watermark could be extracted by statistics concept with the corrupted embedded threshold which could resist the plenty attacks. It reaches the robustness of our design goal.
The third and fourth schemes are without the original image. They are proposed to improve the first and second schemes. The third scheme combines the first and second schemes. The main idea to match the purpose without original image is that the threshold is embedded into three blocks other than the block embedded the watermarking. On the other hand, the fourth and second schemes use the similar concept. The main difference is that the third and fourth schemes would sacrifice a little the imperceptibility of watermarked image and the robustness of watermark to match the purpose without the original image.
The contributions of our research are as follows:
(1) Derive a high-speed watermarking system.
(2) Reach a very well robustness property.
(3) Extract watermarking without original image.
They are described as follows:
In recent years, the proposed schemes by most of the scholars were based on frequency-domain. The main reason is that it is easier to reach imperceptibility, robustness and without the original image than on spatial-domain, simultaneously. But it has a disadvantage that the operation speed of the whole system is slower than on spatial-domain. Therefore, we propose spatial-domain watermarking techniques to derive a high-speed watermarking system with imperceptibility, robustness and without original image.
In the past, for spatial-domain, the proposed schemes by most of the scholars were with the original image. Our first and second schemes are with the original image, too. However, our proposed schemes are more robustness than traditional watermarking schemes. In addition, so far, only few of scholars proposed schemes without original image based on Spatial-Domain. However, the robustness of their proposals was poor. In this thesis, the third and fourth proposals without original image by the authors have very well robustness under some standard attacks. The purpose of without original image is that the copyright owner cannot spend the extra disk space to preserve the original image.
From this study, we believe that the research results will be probably practical and efficient on applications of copyright protection.
|
author2 |
Shih-Hsiung Twu |
author_facet |
Shih-Hsiung Twu Ming-Hui Ni 倪明輝 |
author |
Ming-Hui Ni 倪明輝 |
spellingShingle |
Ming-Hui Ni 倪明輝 A Spatial Robust Watermarking Techniques Based on Image Statistics |
author_sort |
Ming-Hui Ni |
title |
A Spatial Robust Watermarking Techniques Based on Image Statistics |
title_short |
A Spatial Robust Watermarking Techniques Based on Image Statistics |
title_full |
A Spatial Robust Watermarking Techniques Based on Image Statistics |
title_fullStr |
A Spatial Robust Watermarking Techniques Based on Image Statistics |
title_full_unstemmed |
A Spatial Robust Watermarking Techniques Based on Image Statistics |
title_sort |
spatial robust watermarking techniques based on image statistics |
publishDate |
2004 |
url |
http://ndltd.ncl.edu.tw/handle/72863632759462652394 |
work_keys_str_mv |
AT minghuini aspatialrobustwatermarkingtechniquesbasedonimagestatistics AT nímínghuī aspatialrobustwatermarkingtechniquesbasedonimagestatistics AT minghuini yǐyǐngxiàngtǒngjìzīxùnwèijīchǔzhīqiángjiànxìngkōngjiānfúshuǐyìnjìshù AT nímínghuī yǐyǐngxiàngtǒngjìzīxùnwèijīchǔzhīqiángjiànxìngkōngjiānfúshuǐyìnjìshù AT minghuini spatialrobustwatermarkingtechniquesbasedonimagestatistics AT nímínghuī spatialrobustwatermarkingtechniquesbasedonimagestatistics |
_version_ |
1718159232991232000 |