الفهرس | Only 14 pages are availabe for public view |
Abstract This thesis is concerned with a new branch in image authentication, which is Copy-Move forgery detection. The problem of image tampering has spread in the last decade due to the advances in digital image processing tools. This led to different types of image tampering. Of such types is the image Copy-Move forgery, which depends on copying certain parts or objects of images and pasting them in other positions in the same images. This action leads to the creation of fake versions of images. The objective of this thesis is to present efficient techniques for Copy-Move forgery detection. Two techniques are presented for this propose. The first one depends on trigonometric and wavelet transforms with automatic classifications tools. Simulation results of this technique with different alternatives reveal high success rates of Copy-Move forgery detection. In the second technique, the Speeded Up Robust Features (SURF) transform is used for Copy-Move forgery detection. It depends on estimating the feature points of image blocks and taking the number of matched keypoints between each block and those of its complementary image. A correlation matrix is estimated for the number of matched points between blocks. The Singular Value Decomposition (SVD) of this correlation matrix is estimated and the singular values are used for classification of forgery or not. High classification rates have been obtained with this technique. |