الفهرس 
Ch.1: IntroductionOnly 14 pages are availabe for public view
 |  | from | 25 |  |  |
| | | from | 25 | | |
Abstract
Ultrasound imaging has been widely accepted as an essential safe tool for biological tissue characterization and medical diagnosis. Those images are generally affected by speckle noise which is mainly due to the scattering phenomenon‘s coherent nature. The noise filtration-which causes losses of diagnostic features- is treated with introducing a wavelet, Curvelet, and Fusion based image processing used for removing speckles while keeping the fine features of the scan‘s image; enhancing the quantitative tissue characterization of ultrasound imaging. Performance evaluation of our work is done quantitatively by four measures including; the peak signal to noise ratio(PSNR), the square root of the mean square of error (RMSE), a universal image quality index (Q), and the Pratt‗s figure of merit (FOM) as a measure for edge preservation. Plus canny edge map which is extracted as a qualitative measure of edge preservation. In our thesis, double thresholding segmentation after denoising in Curvelet transform domain is proved to be useful for sonogram quantitative diagnosis and measurements for the cases of fetal congenital anomalies-tricuspid valve dysplasia, hepatic abscess detection, and intraocular examination for eyes. This method (which is an enhanced compression algorithm) has the advantages of large reduction in image memory storage area and its processing time. An original Gray phantom is designed to support the validation of our applied enhancement method for gray-scale biomedical Ultra-sonograph diagnosis.
ii
A general quality optimization index S&M (S. Badawy and M. GadAllah) is newly introduced, into image processing research, for selecting the best parameter’s value for image fusion methods being firstly introduced. A case study has been done to characterize the hepatic tissue to obtain a quantitative differentiation between the normal and abnormal Liver. A five quantitative measures were used for the image texture analysis; Entropy, Standard Deviation (SD), GLCM-Contrast (Cont.), GLCM-Correlation (Corr.), and GLCM-Energy. GLCM is stated for Gray-Level Co-Occurrence Matrix. The measurements of Wavelet Image Fusion after Curvelet Denoising, has better speckle Reduction and Edge Enhancements on ultrasound scans, than only denoising based approaches; which lead to better quantitative tissue characterization. The applied approach showed qualitative and quantitative success on image denoising while maintaining edges’ information as possible. The specialist clinical decisions assured the success of the applied processing procedure into producing images having better resolution and better differentiation of some organ‘s texture like kidney, in our case. While less success was noticed for some cases.