Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain

Digital image watermarking techniques have enabled imperceptible information in images to be hidden to ensure the information can be extracted later from those images. Robustness, imperceptibility, capacity and security are the most important requirements of any watermarking scheme. Recently, hybrid...

Full description

Saved in:
Bibliographic Details
Main Author: Mohamed Hassan Makbol, Nasrin
Format: Thesis
Language:English
Published: 2015
Subjects:
Online Access:http://eprints.usm.my/41123/1/NASRIN_MOHAMED_HASSAN_MAKBOL_24_Pages.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
id my-usm-ep.41123
record_format uketd_dc
spelling my-usm-ep.411232018-07-20T07:27:38Z Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain 2015 Mohamed Hassan Makbol, Nasrin TK7800-8360 Electronics Digital image watermarking techniques have enabled imperceptible information in images to be hidden to ensure the information can be extracted later from those images. Robustness, imperceptibility, capacity and security are the most important requirements of any watermarking scheme. Recently, hybrid Singular Value Decomposition (SVD)- based watermarking schemes in the wavelet domain have significantly gained a lot of attention. The aim of this study is to develop hybrid digital image watermarking schemes by combining the properties of SVD and the chosen wavelet transforms to achieve high robustness and imperceptibility, as well as maintaining the trade-off between robustness, imperceptibility and capacity. The security issue due to the false positive problem (FPP) that may be occurring in most of SVD-based watermarking schemes, has been covered and addressed. This study proposes five hybrid robust SVD-based image watermarking schemes in the wavelet domain. In the first scheme, a grey image watermark is embedded directly into the singular values (S) of each redundant discrete wavelet transform transform (RDWT) sub-band of the host image. The scheme is named RDWT-SVD. The second proposed scheme, namely IWT-SVD-AT, utilised the integer wavelet transform (IWT) instead of RDWT due to its properties. The watermark is scrambled using Arnold Transform (AT) before being embedded into the S of each IWT sub-band host. Despite the impressive results by the first and the second schemes, they were vulnerable to the FPP. Thus, they have failed to resolve the rightful ownership. In the third scheme, a hybrid IWT-SVD scheme is proposed with a novel Digital Signature (DS)-based authentication mechanism to solve the FPP. The scheme outperforms the previous schemes in terms of robustness, capacity, security, computation time and attains high imperceptibility. In the remaining two proposed schemes; the fourth and fifth schemes, the FPP is totally avoided using new different embedding strategies. In the fourth scheme namely IWT-SVD-MOACO, the singular vector U of the watermark is embedded into the S of IWT LL sub-band. Multi-objective ant colony optimisation (MOACO) is used to find the optimal multiple zooming/scaling factor (MZF) instead of the single scaling factor (SSF) to achieve the optimal trade-off between imperceptibility and robustness. Finally, a hybrid SVD block-based scheme namely DWT-SVD-HVS using discrete wavelet transform (DWT) is developed. A binary watermark is embedded into a number of blocks which is selected based on some human visual system (HVS) criterion. The scheme shows a high imperceptibility and good robustness. Finally, all the proposed schemes are evaluated with different colour images and had been shown a successful applicability with colour images. 2015 Thesis http://eprints.usm.my/41123/ http://eprints.usm.my/41123/1/NASRIN_MOHAMED_HASSAN_MAKBOL_24_Pages.pdf application/pdf en public phd doctoral Universiti Sains Malaysia Pusat Pengajian Kejuruteraan Elektrik dan Elektronik
institution Universiti Sains Malaysia
collection USM Institutional Repository
language English
topic TK7800-8360 Electronics
spellingShingle TK7800-8360 Electronics
Mohamed Hassan Makbol, Nasrin
Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
description Digital image watermarking techniques have enabled imperceptible information in images to be hidden to ensure the information can be extracted later from those images. Robustness, imperceptibility, capacity and security are the most important requirements of any watermarking scheme. Recently, hybrid Singular Value Decomposition (SVD)- based watermarking schemes in the wavelet domain have significantly gained a lot of attention. The aim of this study is to develop hybrid digital image watermarking schemes by combining the properties of SVD and the chosen wavelet transforms to achieve high robustness and imperceptibility, as well as maintaining the trade-off between robustness, imperceptibility and capacity. The security issue due to the false positive problem (FPP) that may be occurring in most of SVD-based watermarking schemes, has been covered and addressed. This study proposes five hybrid robust SVD-based image watermarking schemes in the wavelet domain. In the first scheme, a grey image watermark is embedded directly into the singular values (S) of each redundant discrete wavelet transform transform (RDWT) sub-band of the host image. The scheme is named RDWT-SVD. The second proposed scheme, namely IWT-SVD-AT, utilised the integer wavelet transform (IWT) instead of RDWT due to its properties. The watermark is scrambled using Arnold Transform (AT) before being embedded into the S of each IWT sub-band host. Despite the impressive results by the first and the second schemes, they were vulnerable to the FPP. Thus, they have failed to resolve the rightful ownership. In the third scheme, a hybrid IWT-SVD scheme is proposed with a novel Digital Signature (DS)-based authentication mechanism to solve the FPP. The scheme outperforms the previous schemes in terms of robustness, capacity, security, computation time and attains high imperceptibility. In the remaining two proposed schemes; the fourth and fifth schemes, the FPP is totally avoided using new different embedding strategies. In the fourth scheme namely IWT-SVD-MOACO, the singular vector U of the watermark is embedded into the S of IWT LL sub-band. Multi-objective ant colony optimisation (MOACO) is used to find the optimal multiple zooming/scaling factor (MZF) instead of the single scaling factor (SSF) to achieve the optimal trade-off between imperceptibility and robustness. Finally, a hybrid SVD block-based scheme namely DWT-SVD-HVS using discrete wavelet transform (DWT) is developed. A binary watermark is embedded into a number of blocks which is selected based on some human visual system (HVS) criterion. The scheme shows a high imperceptibility and good robustness. Finally, all the proposed schemes are evaluated with different colour images and had been shown a successful applicability with colour images.
format Thesis
qualification_name Doctor of Philosophy (PhD.)
qualification_level Doctorate
author Mohamed Hassan Makbol, Nasrin
author_facet Mohamed Hassan Makbol, Nasrin
author_sort Mohamed Hassan Makbol, Nasrin
title Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
title_short Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
title_full Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
title_fullStr Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
title_full_unstemmed Improvement Of Hybrid Digital Image Watermarking Schemes Based On Svd In Wavelet Transform Domain
title_sort improvement of hybrid digital image watermarking schemes based on svd in wavelet transform domain
granting_institution Universiti Sains Malaysia
granting_department Pusat Pengajian Kejuruteraan Elektrik dan Elektronik
publishDate 2015
url http://eprints.usm.my/41123/1/NASRIN_MOHAMED_HASSAN_MAKBOL_24_Pages.pdf
_version_ 1747820878020739072