Image denoising project

Harbir Antil, Assistant Professor, Mathematics, GMUProject description:

Removing noise from an image, or more generally image smoothing, is a fundamental image processing problem with important applications and a vast existing literature. Earliest and most common approaches apply discrete filters to image pixels to achieve a smoothing effect. These filters can often be viewed as discretizations of partial differential equations (PDEs) that would be apply the same smoothing effect in the continuous realm. Conversely, modeling based on partial differential equations can be a very powerful approach to design new image denoising algorithms. This has led to very effective image denoising algorithms using nonlinear PDEs, such as total variation denoising, in the last few decades. Most recently a nonlocal/fractional PDE model has been proposed for this purpose, and has been shown to have distinct advantages. The project will explore the cutting edge denoising techniques in various application areas.