الفهرس 
AbstractOnly 14 pages are availabe for public view
 |  | from | 158 |  |  |
| | | from | 158 | | |
Abstract
Wavelet theory attracted attention in the 1980 through the work of several researchers from various disciplines for various fields includes, human vision, earthquake prediction, pure mathematics, and image processing. This work investigates the use of wavelet transform in image processing. Various applications have benefited from the wavelet, in this work, we describe two of such applications, Denoising and Registration. Noise can be introduced into digitized image in many ways, starting with the lens of imaging hardware and ending at the digitization of the captured image. The reduction of noise without degradation of the underlying images has attracted our attention . In this work, two proposed denoising algorithms based on spatial and wavelet transform domains are represented. The first denosing algorithm in spatial domain based on the local adaptive window size and local information measure to determine weather a pixel is noisy before applying filtering unconditionally. The algorithm constitute a new efficient method of noise reduction in spatial domain. The second densoing algorithm in wavelet transform domain based on multiresolution local contrast entropy of wavelet coefficients. Depending on the distribution of the noise in the wavelet, a new adaptive threshold estimation algorithm is introduced. This threshold enables the proposed algorithm to be adaptive to unknown smoothness of the denoising images. On the other hand, the selection of ground control points play an important role to achive the accuracy of image registration process. This thesis introduce a new approach for automatic control point selection. It looks at the evolution of the coefficients across the different scales and extracts more significant points according to its contribution weight to the multiresolution local contrast entropy.