Noise-Driven Anisotropic Diffusion Filtering for X-Ray Low Contrast Imaging

Abstract

We describe theoretically a possible application of advanced noise-filtering technique to X-ray imaging of low contrast objects, using the example of X-ray studies of laser fusion capsules. Based on the concept of the nonlinear anisotropic diffusion equation, we develop a computer code which uses noise-driven parameterization to optimize the image denoising process. The computation results demonstrate an effective noise filtering while preserving the edge pattern of noisy images obtained with the original contrast as low as ~1%.