Growth of the Internet in the past years has rapidly increased the availability of digital images, video and audio to the public. The availability and efficiency of global computer networks for the communication of digital data has speed up the popularity of digital media. Creative approaches to store, access and distribute data have generated many benefits for the digital multimedia field, mainly due to properties such as free distortion transmission, compact storage, and easy editing. The great facility in copying a digital content rapidly and perfectly without limitations has caused to the problem of copyright protection. Free access digital multimedia communication provides virtually unprecedented opportunities to copy illegally copyright material. The audio-image and video industry has been the victim of such illegal reproduction and consequently, the design and development of effective digital multimedia copyright protection methods have become needed more than ever.
Text watermarking, image watermarking, video watermarking, audio watermarking are the regular applications in audio-visual industry. Basically digital watermarking can be divided into two classes: visible watermarking and invisible watermarking.
Get your grade
or your money back
using our Essay Writing Service!
Visible watermarking technique is usually used by various companies to indicate their ownership. Usually those authorities will insert their logos as the watermark on the digital multimedia content such as videos or images. Visible watermarking technique is widely used in many applications, but nowadays much software is created to remove those visible watermarking especially image editing software such as Adobe Photoshop. Watermark may lose its functions to show the copyright of the digital products with the appearance of this software. Therefore invisible watermark also needed in this market.
Invisible watermarking technique is a different technique used to embed certain information or message in the products. The main difference compared to visible watermarking is the watermark will be hidden inside the product and cannot be identified by naked eyes. The only way to get back the information or hidden message is to use certain algorithm to extract or detect the watermark. Others will not be able to detect the hidden information without the right key.
Figure 1 (a): Original Lena Image
Figure 1 (b): Example of visible watermarking; (c) Invisible watermarking
Figure 1 illustrates the difference between visible watermarking and invisible watermarking technique.
Digital watermarking is proposed as a solution to this challenging issue to prove the ownership of digital data. A watermark can be a random signal, a significant message or a company's logo that can be used for ownership protection, copy control and authentication. Only the authorized users should gain access to the watermarked data. In order for a digital watermark to be effective, it should be robust to common image manipulations such as compressions, filtering, rotation, Attackers have the freedom to obtain copies of copyrighted multimedia information via Internet, so protection of digital multimedia information has attracted many attentions.
Ton Kalker defines watermarking security as "the inability by unauthorized users to have access to the raw watermarking channel". In other words, watermark security refers to the failures of unauthorized users to alter, to remove, to read or to write the watermark content established by robust watermarking. Security deals with intentional and unintentional attackers whose aims to eliminate, remove or degrade the effectiveness of the watermark  .
The idea of robust watermarking is to embed watermarked data within the image with an insensitive form for human visual system but in a way to overcome the attacks from common image processing or intentional attempts to remove or alter it. The aim is to produce an image that looks exactly the same to human eyes without affecting its visual content.
A general scheme for digital watermarking is given in Figure 2.
Figure 2: A general digital watermarking system
A watermark message W is embedded into an original image H by using a secret key at the coder. Only the owner of the data knows the secret key. It is hard to remove or alter the message from the data without knowing the key.
The resulting image is the watermarked image H*. It is passes through the transmission channel. The digital watermarked product will be transmitted through some ways such as internet, or transmission within pen-drive.
Always on Time
Marked to Standard
The transmission channel includes the possible attacks, such as lossy compression, any signal processing operation, geometric distortions and etc.
After the watermarked image passes through all these operations, the message is tried to be detected or extracted at the decoder. The extracted watermark will be compared with the original watermark to prove that the copyright of current products are belongs to right people.
Even though digital watermarking can be applied in many digital data such as video, audio or image, but this project will scope into image digital watermarking and the watermark used will also be in image type.
2.0 Problem Statement
Computer technology allows people to duplicate easily and distribute digital multimedia through the Internet. However, these benefits bring risk of data piracy. This may lead to serious piracy problem because digital multimedia editing software for audio, document, image and video is widely used and widely spread through internet. Internet users can just simply download those digital images and remove visible watermarks on images easily. The digital images will lose their copyright after edited. One solution to provide security for copyright protection is robust watermarking, which ensures that embedded messages survive from various watermarking schemes such as image processing attacks, geometric transformation, cryptographic attacks, and protocol attacks. If the watermarked message is under an attack, the hidden data may be destroyed and the recipient cannot identify the correct message.
3.0 Literature Review
Vallabha VH defines watermarking as "hiding of a secret message or information within an ordinary message and the extraction of it at its destination. In other words, watermarking is a process of embed hidden data called watermark message into the original message. The watermark message should be detectable or recoverable from the watermarked image even though it has been altered by various watermarking attack methods.
3.2 Classification based on characteristics
Digital watermarking can be categorized into five classes according to the characteristics of embedded watermarks  :
Blind versus nonblind
Perceptible versus imperceptible
Private versus public
Robust versus fragile
Spatial domain versus frequency domain
3.2.1 Blind versus nonblind
Blind technique does not require access to the original data such as audio, image; video etc recover the watermark image. Conversely, the original data are needed for the extraction of the watermark image for the nonblind technique. Technique that is relying on the original data for watermark detection is usually more robust but the limit they require is very strong. Typically, the nonblind technique is more robust than the blind technique because the watermark message can be extracted easily by knowing the unwatermarked data. However, mostly in real applications the unmodified host signal is not available to the watermark detector. The blind technique does not need the original data, thus it is more useful than the nonblind technique in most applications.
3.2.2 Perceptible versus imperceptible
A watermark is classified as a perceptible if the embedded watermark is planned to be visible. A good perceptible watermark must difficult to be removing or altered by unauthorized person. Conversely, an imperceptible watermark is embedded into a host image by sophisticated algorithms .Plus; it is invisible to the naked eye. However, the watermark can be extracted by a computer.
3.2.3 Private versus public
A watermark is said to be private if only authorized person can access it. Only those who have the access can make any changes of the watermark. Thus, private watermarking need a security key to protects from unauthorized person to extract the watermark. This private security key specifies a watermark's location in the host image, and allows any insertion and removal of the watermark if the secret location is identified.
On the other hand, watermarking techniques that allows anyone to access the watermark are called public. Public watermarks are embedded in a location known to everyone. Thus, the watermark detection software can easily extract the watermark by scanning the whole images. As a consequence, private watermarking is more robust compared to the public watermarking.
3.2.4 Robust versus fragile
Watermark robustness explains for the capability of the hidden watermark to survive legal of daily usage or any image processing manipulation from intentional and unintentional attackers. Means that, robust watermark has the availability to withstand various image attacks thus providing authentication.
This Essay is
a Student's Work
Intentional attacks refer to the people who aim at destroying the watermark, while unintentional attacks do not clearly plan to alter it.
According to the embedding purposes in term of fidelity, watermarks can be classified in three types: robust, semi fragile and fragile.
Robust watermark: Designed to survive from intentional and unintentional modifications of the watermarked image. It is usually used for copy right protection to declare rightful ownership.
Semi fragile watermark: This kind of watermark is intermediate between fragile and robust watermark. It is designed for detecting any unauthorized modifications and together allowing some image processing operations. Semi fragile watermark are commonly used for selective authentication that detects illegitimate distortion while ignoring applications of legitimate distortion.
Fragile watermark: It has the ability to withstand various image attacks and also provide s authentications. Fragile watermarking techniques are concerned with complete reliability verification. The slightest modification of the watermarked image will alter and can also destroy the fragile watermark.
3.2.5 Spatial domain versus frequency domain
Additional information called watermark is embedded into the original data in such a way that it remains present since the perceptible quality of the content is at an acceptable level. The owners of the original data prove their ownership by extracting the watermark from the watermarked content.
The image watermarking algorithms can be classified into two categories: the spatial domain and the frequency domain.
In the spatial domain, we can simply insert a watermark into a host image by changing its pixels in the host image. Spatial domain has the advantages of low complexity and easy implementation. However, the inserted information may be easily detected using computer analysis and can be easily attacked.
In the frequency domain, the watermark is embedded into the coefficients of a transformed image such that the watermark is invisible and more robust for some image processing operations. This method separates the watermark in spatial domain of the image, hence making it very hard to remove the embedded watermark. Frequency domain techniques proved to be more effective with respect to achieving the imperceptibility and robustness requirement in digital watermarking algorithms. The transformations include discrete cosine transform (DCT), discrete wavelets transform (DWT) and discrete Fourier transforms (DFT).
3.3 A General Watermarking Structure
A complete watermarking process consists of embedding and detection parts. Theoretically, the components of a watermark embedding/detection/extraction system are shown in Figure 3  .
Figure 3: A general watermarking scenario
As shown in Figure 3, the watermark generator takes an original image and a key to generate the watermark.
The encoder then embeds the watermark into original image to produce a watermarked image. The watermarked image is then distributed and potentially distorted by image processing operations and targeted watermark attacks. Lastly, the decoder attempts to find out the existence of the original watermark in the distorted image.
The embedded data can be detected or extracted from digital multimedia element in an application. Important properties of an image watermarking system include perceptible transparency, robustness, and capacity.
Perceptible transparency refers to the systems that use characteristics of the human visual system (HVS) to assure that the watermark is invisible under usual viewing situations. In other words, a watermarked image should not see any different from the original image.
Robustness is the resistance of the watermark against intentional attacks such as addition of noise, filtering, lossy compression, scaling, rotation and etc. Watermark image still can be detected after the image has undergone various attacks from malicious person or not.
Data capacity refers to the amount of data that can be embedded without affecting perceptual transparency. The relative importance of these properties depends on the requirements of a given application. Its allow insertion of multiple, independently detectable watermarks in an image.
A proper selection of the frequency domain is dependent on the application. The better the image transform approximates the properties of the human visual system (HVS), the easier is to put more energy in the embedded signal without causing perceptible distortion. According to the HVS, the high frequencies are less visible than the low frequencies. However, if we embed too much data in the frequency domain, the image quality will be degraded significantly. In this project, a robust watermarking technique based on DWT is proposed because it is more close to the HVS than DCT.
3.4 Proposed Watermarking Technique
The watermark embedding can be either is spatial domain of frequency domain of an image. Spatial domain watermarking schemes are usually more efficient than the frequency domain, but the frequency domain watermarks are more resistant to various watermark attacks. As s result, almost all recently research proposed watermarking schemes operates in a transform domain.
Discrete Wavelet Transform (DWT) has been used in digital watermarking more frequently compared to other types of frequency domain as mentioned before. This is due to its excellent spatial localization of most important image features such as textures and edge. The wavelet transform has multi resolution characteristics of images. It benefits progressive transmission of image data and hierarchical decoding of nested watermarks. In addition, The DWT provides better modeling of the HVS because the wavelet transforms match the multi channel model of the HVS. This advantage of DWT allows using higher energy watermarks in regions where the HVS is less sensitive, so that embedding watermarks in these regions provides to increase the robustness of the watermarking technique  .
Therefore, in this paper, digital watermarking technique based on DWT is proposed. No research regarding digital watermarking has been conducted in Malaysia especially in image industry. Probably no one realizes about the benefits of watermarking to use for protection the origins of the digital multimedia. Therefore, digital watermarking based on DWT is presented.
3.5 Possible Attacks on Image Watermarking
The attackers intending at the watermarked images can be classifies as unintentional or intentional. The attackers have three strategies to defeat watermark robustness as:
To remove enough watermark signal
To jam the hidden communication channel
To desynchronize the watermarked content
The aim of the attackers is to alter, remove or degrade the effectiveness of the watermark. An attack is said to be successful if the attackers disturb any stage of the watermarking cycle.
For that reason, the owner of digital content and digital watermarking have to make sure that each stage is secured against any image signal manipulations.
Watermarking attacks schemes can be categorized into four classes  :
Image processing attacks
Image processing attacks consist of filtering attacks, remodulation attacks, JPEG coding distortion and JPEG 2000 compression.
Geometric attacks is refers to scaling, rotation, clipping, bending, template, etc. The original image is not needed for extracting the watermarked image.
Cryptographic attacks try to avoid the security of a cryptographic system by retrieving the key in a code or protocol. This type of attacks tries to find out the key used for embedding. Once the key is found, the watermark is overwritten.
Protocol attacks aim to generate protocol uncertainty in the watermarking process. One type of protocol attacks, called Copy Attack aim to estimate a watermark from watermarked data. Then, copy it to some other data instead of destroying the watermark.
The objectives are prepared based on the problem statement mentioned before. The objectives of this work are as follows:
To study the available watermarking techniques
To implement the DWT-based watermarking algorithm that more focused in digital image.
To implement a digital watermarking system on Matlab software using image and Wavelet Toolboxes
To analyze and evaluate the robustness of the proposed watermarking algorithm.
6.0 Contribution to the body of knowledge
Digital watermarking system with embedding and extraction will be created which is friendly-user for those users. Users can use this system to embed the information they want into digital images. Plus, users can extracts the image to show that whether the image contain any information inside this image.