الفهرس 
الغلافOnly 14 pages are availabe for public view
 |  | from | 124 |  |  |
| | | from | 124 | | |
Abstract
Digital images have a great impact on our everyday activities. Due to the availability of many advanced image processing software, it is easy to modify or manipulate digital images without leaving traces of tampering. Therefore, there is a need to check the authenticity and integrity of a given digital image. Copy-move forgery is one of the most common methods of image tampering. Its main purpose is to hide an object or a region in an original image by copying a part of the image and moved it to another location in the same image.
This thesis proposes a methodology for efficient and robust detection of different cases of copy-move forgery. The main purpose of this methodology is that the performance of the detection is gradually enhanced through finding the set of optimum parameter values.
This new approach can efficiently detect both squared and random-shaped tampered regions. It can also detect vertically and/or horizontally flipped tampered regions. Moreover, it can detect tampered region containing holes, intersected regions, and multiple copy-move regions. The processes of adding either white Gaussian noise or blurring are very common in digital image tampering. They could be used to conceal the influence caused by copy-move process and to remove unwanted defects. The experimental results showed that the developed detection algorithm can detect tampered regions that are distorted by Gaussian white noise or blurring. A condition for detection of a random-shaped tampered region is determined. Experiments showed that the accuracy of detection depends on the shape oftampered regions.
Two different detection algorithms are tested to reach the lowest computational time.
The Kd-tree searching algorithm proved to be superior as a block matching step, compared to the lexicographical sorting algorithm. A new algorithm is developed to enhance the performance of detection by eliminating the false-negative regions. A user¬friendly graphical user interface is built to show the results.