Face Image Denoising Method Based on Fourth-Order Partial Differential Equations

In order to improve the effect of face image denoising, this paper put forward several face image denoising methods based on partial differential equations, including P-M non-linear diffusion equations and fourth-order partial differential equations. We use those two methods by establishing non-linear diffusion equations and fourth-order anisotropic diffusion partial differential equation. The P-M non-linear diffusion denoising method can remove noise in intra-regions sufficiently but noise at edges can not be eliminated successfully and line-like structures can not be held very well.While the fourth-order partial differential equations denoising can retain the local detail characteristics of the original face image. Finally, through the experimental results we can see the effect of the fourth-order partial differential equations denoising is better, which makes the later face image processing more accurate and promotes the development of face image processing.

Download Full-text

Applying Fourth-Order Partial Differential Equations and Contrast Enhancement to Fluorescence Microscopic Image Denoising

Advances in Intelligent and Soft Computing - Knowledge Discovery and Data Mining ◽

10.1007/978-3-642-27708-5_17 ◽

2012 ◽

pp. 123-128 ◽

Cited By ~ 8

Author(s):

Yu Wang ◽

Hong Xue

Keyword(s):

Partial Differential Equations ◽

Differential Equations ◽

Contrast Enhancement ◽

Image Denoising ◽

Fourth Order ◽

Microscopic Image ◽

Partial Differential

Download Full-text

Correlation Moment Analysis and the Time Dependence of Coherence in Systems Described by Nonlinear Partial Differential Equations

Journal of Applied Mechanics ◽

10.1115/1.2065687 ◽

2005 ◽

Vol 73 (2) ◽

pp. 197-205 ◽

Cited By ~ 1

Author(s):

M. R. Belmont

Keyword(s):

Partial Differential Equations ◽

Differential Equations ◽

Time Dependence ◽

Initial Conditions ◽

Nonlinear Partial Differential Equations ◽

Moment Analysis ◽

Partial Differential ◽

Linear Diffusion ◽

Linear Partial Differential Equations ◽

Non Linear

The work presented introduces correlation moment analysis. This technique can be employed to explore the growth of determinism from stochastic initial conditions in physical systems described by non-linear partial differential equations (PDEs) and is also applicable to wholly deterministic situations. Correlation moment analysis allows the analytic determination of the time dependence of the spatial moments of the solutions of certain types of non-linear partial differential equations. These moments provide measures of the growth of processes defined by the PDE, furthermore the results are obtained without requiring explicit solution of the PDE. The development is presented via case studies of the linear diffusion equation and the non-linear Kortweg de-Vries equation which indicate strategies for exploiting the various properties of correlation moments developed in the text. In addition, a variety of results have been developed which show how various classes of terms in PDEs affect the structure of a sequence of correlation moment equations. This allows results to be obtained about the behavior of the PDE solution, in particular how the presence of certain types of terms affects integral measures of the solution. It is also demonstrated that correlation moments provide a very simple, natural approach to determining certain subsets of conserved quantities associated with the PDEs.

Download Full-text