Detail publikace

Using Partial Differential Equations to Filter Images with High Noise Levels

Originální název

Using Partial Differential Equations to Filter Images with High Noise Levels

Anglický název

Using Partial Differential Equations to Filter Images with High Noise Levels

Jazyk

en

Originální abstrakt

The paper discusses image noise suppression methods based on the solution of partial differential equations. The individual techniques are applied to a set of test images and then compared with respect to the obtained noise suppression levels. One of the presented procedures is image denoising via the heat conduction formula, which constitutes an equivalent of the Gaussian filter. Further, a bilateral filter is derived; this approach better preserves the edges of objects within the examined image. The last technique analyzed is anisotropic diffusion (Perona-Malik). Importantly, the signal-to-noise ratio values obtained with these different filtering methods are compared in the final section of the article.

Anglický abstrakt

The paper discusses image noise suppression methods based on the solution of partial differential equations. The individual techniques are applied to a set of test images and then compared with respect to the obtained noise suppression levels. One of the presented procedures is image denoising via the heat conduction formula, which constitutes an equivalent of the Gaussian filter. Further, a bilateral filter is derived; this approach better preserves the edges of objects within the examined image. The last technique analyzed is anisotropic diffusion (Perona-Malik). Importantly, the signal-to-noise ratio values obtained with these different filtering methods are compared in the final section of the article.

BibTex


@inproceedings{BUT136331,
  author="Jiří {Sliž}",
  title="Using Partial Differential Equations to Filter
Images with High Noise Levels",
  annote="The paper discusses image noise suppression methods based on the solution of partial
differential equations. The individual techniques are applied to a set of test images and
then compared with respect to the obtained noise suppression levels. One of the
presented procedures is image denoising via the heat conduction formula, which
constitutes an equivalent of the Gaussian filter. Further, a bilateral filter is derived;
this approach better preserves the edges of objects within the examined image. The
last technique analyzed is anisotropic diffusion (Perona-Malik). Importantly, the
signal-to-noise ratio values obtained with these different filtering methods are
compared in the final section of the article.",
  address="VUT FEKT UTEE",
  booktitle="International Interdisciplinary PhD Workshop 2016 Proceedings",
  chapter="136331",
  howpublished="electronic, physical medium",
  institution="VUT FEKT UTEE",
  year="2016",
  month="september",
  pages="82--86",
  publisher="VUT FEKT UTEE",
  type="conference paper"
}