الفهرس Only 14 pages are availabe for public view
 |  | from | 140 |  |  |
| | | from | 140 | | |
Abstract
Images are an important tool in many different fields. With the rapid increase in the use of computers, the images are now used in digital form. Digital images may be distorted during capturing, digitization, or transmission processes. Image enhancement and restoration operations concern with removing or reducing these distortions. In this work, image denoising problem is investigated. Image denoising task involves the manipulation of the image data to produce a visually high quality image. This thesis reviews some of the standard previously published denoising algorithms, such as filtering approaches, wavelet based approaches, fractal based approaches, and fractal-wavelet (FW) approaches. Its aim is to develop and experiment an efficient hybrid fractal-wavelet coder, which can be applied for image denoising. Fractal-wavelet coding schemes provide a better reconstructed image quality and a faster decoding process than fractal methods. The problem with FW schemes is their dependence on the chosen resolution. For a low resolution, the FW encoding methods are fast but yield overly denoised representation at the expense of smoothing the edges of the image and creating artificial ringing artifacts. On the other hand, for a high resolution, FW coding is slower but of a better quality of the noisy image. Nevertheless, it does not perform enough denoising, resulting in sharp but visibly noisy representations. The proposed technique presented in this thesis is designed to take the advantage of two of the famous FW coding schemes, the standard FW scheme and the generalized FW scheme, in order to overcome the drawbacks of using low resolution in FW coding. The proposed composite technique has been successfully implemented and tested. Comparative results are promising.