Saeed AL-Mansoori, Assistant Research Engineer, Emirates Institution for Advanced Science & Technology (EIAST) has developed a technique to protect the intellectual property of DubaiSat-1.
The aim was to implement a robust algorithm to hide the EIAST logo within any delivered image in order to increase the data security and protect the ownership of DubaiSat-1 images. At the same time, it was necessary to provide high quality images to the end-user
Data security is critical in the field of remote sensing,” says Saeed AL-Mansoori, Assistant Research Engineer, Emirates Institution for Advanced Science & Technology (EIAST).
Currently pursuing an M.Sc. degree in Electrical Engineering from the American University of Sharjah, he undertook, in the beginning of 2012, to explore an effective way to protect the intellectual property of images acquired through Dubai’s earth observation satellite – DubaiSat-1 – and the subsequent satellites. In June 2012, the published a study titled: ‘An efficient watermarking technique for satellite images using Discrete Cosine Transform’.
Giving us an overview of the background and significance of the study, Al-Mansoori states: “To deal with the important issues of rights protection and authentication of images for EIAST, the ‘watermarking’ concept was introduced. The idea of digital image watermarking is to embed the information within a signal (i.e. image, video, etc.), in a manner that cannot be easily extracted by a third party.
[boxposthighlight]The threats
Noise addition attacks
Noise addition is one of the attacks used to test the robustness of watermarking techniques. Basically, by adding noise to an image, invisible data within the image will be destroyed since unnecessary data will replace the image data.
Cropping attacks
The cropping attack removes some components of the watermarked image.
Rotation attacks
In digital multimedia images, “Rotation Attack” is considered as the most popular kind of geometrical attack. The concept behind this attack is to rotate the watermarked image clockwise by a certain degree. This will affect the watermark within the image.
JPEG compression attacks
This attack is considered as one of the most severe attacks on watermarked images. Al Mansoori and his team tested the robustness of the watermark as six levels of JPEG compression quality factors were implemented.
[/boxposthighlight]“The aim was to implement a robust algorithm to hide the EIAST logo within any delivered image in order to increase the data security and protect the ownership of DubaiSat-1 images. At the same time, it was necessary to provide high quality images for the end-user; and adding any information (logo) to these images will affect its quality. Therefore, a model of watermarking had to be designed that would keep the observable difference between the watermarked and original image to a minimum.”
It was important to address the twin challenges of protecting copyright while not compromising on the quality of the image for the end-user.
Al Mansoori explains: “It was crucial to ensure that the watermark should be difficult to remove or alter without the degradation of the host image. This study was based on the Discrete Cosine Transform (DCT) approach to provide highly robust protection in cases such as noise addition, cropping, rotation and JPEG compression attacks.” (See box titled ‘The threats’)
Saeed Al-Mansoori and his co-author (see below) decided to adopt the invisible watermarking approach.
“Invisible watermarking is a digital process wherein the watermark is not perceptible, but may be extracted by computer software.
Most of recent studies are focused on this type of watermarking.”
Watermarking requirements
In order to have a functioning watermarking system, a number of requirements must be achieved;
(a) Security: Watermarking security means that the system is difficult to remove or change/ modify without destroying the original (host) signal. The main idea of any watermarking system is to protect the information embedded without losing the content of the signal to assure secrecy of information.
(b) Imperceptibility: Ideally, there ought to be no observable difference between the watermarked and original image, and the watermark should be difficult to remove or alter without the degradation of the host signal.
(c) Capacity: The term “capacity” in watermarking performance means the maximum percentage of information that
can be hidden in the host image and can be detected when required to prove ownership and authenticity. In general, capacity requirement affects two main parameters; robustness and imperceptibility, that are inversely proportional to each other. As a result, increasing any parameter will affect the other. In case robustness against attacks is required, then a huge amount of information must be embedded which will affect the image quality so that less information can be taken from the watermarked image.
(d) Robustness: The main idea of robustness is that, the hidden watermark must be robust against any attack and image processing transformation which includes filtering, adding noise, rotation, compression and scaling.
The process of embedding
“Discrete Cosine Transform (DCT) is a frequency linear transformation technique used in various applications correlated with a digital signal processing technology,” explains Al-Mansoori. He adds, “Due to its superior performance, it has been used for image compression and pattern recognition.”
In recent years, a significant number of studies have reportedly been done based on the DCT watermarking techniques to implement a robust and secure watermarking scheme. The main advantage is that instead of the researcher selecting the frequency band, the algorithm will randomly select the bands (optimal frequency bands coefficients) to embed the watermark into the original image.
Simulation results and analysis
“The main purpose of this study was to devise a watermark that was robust against several attacks such as JPEG compression, cropping, rotation and noise attacks.
“The methods used for embedding proved their performance, achieving PSNR (Peak signal-to-noise ratio) values ranging from 43.3217 dB to 55.7113 dB. What this means is that, no observable difference emerged between the watermarked and original images.”
“The process can be time-consuming when you consider that satellite images can be as heavy as 12 gigabytes,” reveals Al Mansoori. Armed with the requisite workstation to expedite the process of watermarking, EIAST is slated to implement Al-Mansoori’s recommended method of watermarking in 2013. PRO
About the authors:
Saeed AL-Mansoori is an assistant researcher in the Space Image Processing and Analysis department at the Emirates Institution for Advanced Science and Technology (EIAST). Saeed’s research interests are in the area of image processing (super-resolution, watermarking, object detection and image classification).
Alavi Kunhu Panthakkan is a postgraduate in Digital Electronics Engineering (M.Tech). Alavi works in the Khalifa University in the Electronics Engineering Department as a Lab Instructor.
Add Comment