Doctor of Philosophy (PhD)
Electrical and Computer Engineering
First Committee Member
Emma E. Regentova
Second Committee Member
Third Committee Member
Fourth Committee Member
Ajoy K. Datta
Fifth Committee Member
Number of Pages
Most of digital image applications demand on high image quality. Unfortunately, images often are degraded by noise during the formation, transmission, and recording processes. Hence, image denoising is an essential processing step preceding visual and automated analyses. Image denoising methods can reduce image contrast, create block or ring artifacts in the process of denoising. In this dissertation, we develop high performance non-linear diffusion based image denoising methods, capable to preserve edges and maintain high visual quality. This is attained by different approaches: First, a nonlinear diffusion is presented with robust M-estimators as diffusivity functions. Secondly, the knowledge of textons derived from Local Binary Patterns (LBP) which unify divergent statistical and structural models of the region analysis is utilized to adjust the time step of diffusion process. Next, the role of nonlinear diffusion which is adaptive to the local context in the wavelet domain is investigated, and the stationary wavelet context based diffusion (SWCD) is developed for performing the iterative shrinkage. Finally, we develop a locally- and feature-adaptive diffusion (LFAD) method, where each image patch/region is diffused individually, and the diffusivity function is modified to incorporate the Inverse Difference Moment as a local estimate of the gradient. Experiments have been conducted to evaluate the performance of each of the developed method and compare it to the reference group and to the state-of-the-art methods.
Adaptive; Context; Denoising; Diffusion; Feature; Image processing – Digital techniques; Imaging systems – Image quality; Textons
Computer Engineering | Computer Sciences | Electrical and Computer Engineering
Mandava, Ajay Kumar, "Nonlinear Adaptive Diffusion Models for Image Denoising" (2013). UNLV Theses, Dissertations, Professional Papers, and Capstones. 2007.