M.Ing. / The objective of this research is to develop a new robust and efficient adaptive watermarking method for JPEG compressed images, in which the watermark can survive even if the watermarked image undergoes manipulations. In this thesis, we present two image watermarking methods: one based on the spatial domain and the other on the DCT-based domain. Both approaches, which embed meaningful information into images for copyright protection, are blind watermarking schemes. To illustrate the robustness and performance enhancement derived from Error Correction Codes (ECC) when applied to watermarking, we propose Method I based on the Direct Sequence Spread Spectrum (DSSS) technique in conjunction with Reed-Solomon codes for raw images in the spatial domain. As is evident from our results, the introduction of Reed- Solomon codes improves the robustness of a watermark. Also, employing the properties of pseudo-noise (PN) to deal with cropping attacks is investigated for this method. Based on Method I, we propose Method II that takes into account the properties of JPEG images and exploits the advantages of error correction codes (ECC). Furthermore, some modifications are introduced in the basic Spread Spectrum watermarking to improve the accuracy of the watermark retrieval process. Several experimental results are provided to demonstrate the performance of the proposed schemes. In these experiments, JPEG compression, cropping, brightness/contrast adjustment, median filtering and noise addition are used as watermark attacks to evaluate the robustness. The experimental results show that the proposed scheme (Method II) is very robust against common image processing attacks.

The digital image watermarking is a process that embeds a secrete sequence into a digital image or a video segment to protect its copyright information. There are several methods utilized to deal with the watermarking problem. There are in total two different kinds: the popular neural network and the traditional methods. And for the traditional methods, according to their working region, they can be divided into two groups: spatial domain-related algorithms (e.g., LSB SVD) and transformed domain algorithms (e.g., DCT, DWT). The spatial domain algorithms are the methods that directly work on the pixel values of the image, while the transformed domain algorithms are the methods that work on the rate of change of the image pixels. In the thesis, first of all, we are going to propose a modified hybrid the scheme, which has better results compared with the paper (Lin and Wan 2016). Then, we are going to offer a brand-new method concerning the LDPC-LDGM coding structure in the area of information theory to deal with the digital watermarking problem. Notice that this LDGM-LDPC nested code watermarking scheme only provides a particular case example, which is implemented by one of the teammates Dr. Mahdi, and there is still much space for improvement. However, according to the step tests in chapter 4.4, we have achieved a reasonably good result in imperceptibility and robustness. / Thesis / Master of Applied Science (MASc)

Chen Benrong. / Thesis (M.Phil.)--Chinese University of Hong Kong, 2003. / Includes bibliographical references (leaves 125-130). / Abstracts in English and Chinese. / Chapter 1 --- Introduction --- p.1 / Chapter 1.1 --- What' s Watermarking --- p.1 / Chapter 1.2 --- "Information Hiding, Steganography, and Watermarking" --- p.3 / Chapter 1.3 --- History of Watermarking --- p.5 / Chapter 1.4 --- Importance of Digital Watermarking --- p.8 / Chapter 1.5 --- Objectives of the Thesis --- p.9 / Chapter 1.6 --- Thesis Outline --- p.10 / Chapter 2 --- Applications and Properties of Audio Watermarking --- p.12 / Chapter 2.1 --- Applications --- p.13 / Chapter 2.1.1 --- Ownership Identification and Proof --- p.13 / Chapter 2.1.2 --- Broadcast Monitoring --- p.16 / Chapter 2.1.3 --- Other Applications --- p.18 / Chapter 2.2 --- Properties --- p.19 / Chapter 2.2.1 --- Transparency --- p.20 / Chapter 2.2.2 --- Robustness --- p.20 / Chapter 2.2.3 --- Other Properties --- p.21 / Chapter 3 --- Possible Methods for Audio Watermarking --- p.24 / Chapter 3.1 --- Overview of Digital Audio Watermarking System --- p.25 / Chapter 3.2 --- Review of Current Methods --- p.27 / Chapter 3.2.1 --- Low Bit Coding --- p.27 / Chapter 3.2.2 --- Phase Coding --- p.28 / Chapter 3.2.3 --- Echo Coding --- p.29 / Chapter 3.2.4 --- Spread Spectrum Watermarking --- p.30 / Chapter 3.3 --- Other Related Approaches --- p.31 / Chapter 3.4 --- Outline of Proposed New Method --- p.33 / Chapter 4 --- Audio Watermarking System Based on Spread Spectrum --- p.36 / Chapter 4.1 --- Introduction --- p.36 / Chapter 4.2 --- Embedding and Detecting Information Bit --- p.39 / Chapter 4.2.1 --- General Embedding Process --- p.39 / Chapter 4.2.2 --- General Detection Process --- p.43 / Chapter 4.2.3 --- Pseudorandom Bit Sequences (PRBS) --- p.45 / Chapter 4.3 --- An Optimal Embedding Process --- p.48 / Chapter 4.3.1 --- Objective Metrics for Embedding Process --- p.48 / Chapter 4.3.2 --- Content Adaptive Embedding --- p.52 / Chapter 4.3.3 --- Determination of Frame Length L --- p.57 / Chapter 4.4 --- Requirement For Transparency Improvement --- p.58 / Chapter 5 --- Sample and Frame Selection For Transparency Improvement --- p.60 / Chapter 5.1 --- Introduction --- p.60 / Chapter 5.2 --- Sample Selection --- p.61 / Chapter 5.2.1 --- General Sample Selection --- p.62 / Chapter 5.2.2 --- Objective Evaluation Metrics --- p.65 / Chapter 5.2.3 --- Sample Selection For Transparency Improvement --- p.66 / Chapter 5.2.4 --- Theoretical Analysis of Sample Selection --- p.87 / Chapter 5.3 --- Frame Sclcction --- p.90 / Chapter 5.3.1 --- General Frame Selection --- p.91 / Chapter 5.3.2 --- Frame Selection For Transparency Improvement --- p.94 / Chapter 5.4 --- Watermark Information Retrieve --- p.103 / Chapter 6 --- Psychoacoustic Model For Robustness Verification --- p.105 / Chapter 6.1 --- Introduction of Human Auditory System --- p.106 / Chapter 6.1.1 --- Absolute Hearing Threshold --- p.106 / Chapter 6.1.2 --- Critical Bands --- p.108 / Chapter 6.1.3 --- Masking Effect --- p.111 / Chapter 6.2 --- Psychoacoustic Model of Human Auditory System --- p.112 / Chapter 6.3 --- Robustness Verification by Psychoacoustic Model Analysis --- p.117 / Chapter 7 --- Conclusions and Suggestions For Future Research --- p.121 / Chapter 7.1 --- Conclusions --- p.121 / Chapter 7.2 --- Suggestions For Future Research --- p.123 / Bibliography --- p.125

Digital watermarking technology is a potential future technology, which uses to protect copyright. Traditional digital watermarking is usually suffered from the limited information capacity and imperceptibility deficiency. In this research, we present the method to increase the capacity of the traditional watermarking techniques, which can be used broadly in present digital watermarking techniques by using the relation between the information in the cover-media. We use two kinds of digital watermarking techniques to analysis. The experiment result show that the proposed method can increase the capacity of embedding information without involves significant quality degradation in cover-media.

There has been a growing interest in reversible watermarking of medical images re- cently for security reasons. Typically, humans are assumed to be the end user of watermarked images, however in many cases machine vision processes may be addi- tional consumers. Therefore, any watermarking performed on these images must be imperceptible to not only human users, but also these machine vision processes. The objective of this thesis is to understand the extent to which reversible water- marking affects the ability of computer vision algorithms to perform correctly. We address both the effect on primitive feature detection and on complete machine vi- sion processes, and investigate the ability to predict these effects using Image Quality Metrics (IQMs). Additionally, we describe the development of a new watermarking algorithm. We perform primitive feature detection on original and watermarked images, com- paring the output feature maps. Subsequently we use statistical modelling to allow prediction of feature map differences based on various IQMs of a watermarked image. We then conduct a similar experiment using manually specified feature maps and edge detectors across their full parameter space. Watermarking algorithms showing the least impact are highlighted and prediction of poorer performance a priori is investigated. In many cases watermarking is shown to cause a significant difference in the output feature map, however prediction of the difference is possible with excellent discrim- ination in many cases. A validation system for utilising these results in practical applications is presented. Three machine vision processes are investigated using a range of watermarking al- gorithms and embedding capacities – iris recognition, medical image registration, and diabetic retinopathy assessment. Significant differences are found in some cases, however at low capacities the iris and retinopathy processes show no significant dif- ferences. In addition, prediction of erroneous results for the retinopathy process was possible with excellent discrimination.

Thesis (Ph.D.)--University of Texas at Dallas, 2007. / Includes vita. Includes bibliographical references.

Thesis (M. Phil.)--Hong Kong University of Science and Technology, 2004. / Includes bibliographical references (leaves 70-72). Also available in electronic version. Access restricted to campus users.

Thesis (M.S.)--Rochester Institute of Technology, 2007. / Typescript. Includes bibliographical references (leaves 41-42).

Thesis (Ph. D.)--State University of New York at Binghamton, Watson School of Engineering and Applied Science (Systems Science), 2005. / Includes bibliographical references.

University of Macau / Faculty of Science and Technology. / Department of Computer and Information Science