Publication detail

Generalized Cross Validation with Bayes Approach for Image Denoising

ČERVINKA, T.

Original Title

Generalized Cross Validation with Bayes Approach for Image Denoising

Type

conference paper

Language

English

Original Abstract

This paper describes a new method for suppression of noise in images based on wavelet transform. Classic de-noising methods based on the wavelet transform are based on binary decision. Wavelet coefficients below threshold are replaced by zero, and kept (hard-threshold) or shrinked (soft-threshold) with absolute value above the threshold. The proposed method relies on two criteria. The first criterion is based on estimation of optimal threshold without knowledge exact data using Generalized Cross Validation (GCV) technique. The second criterion employes Bayes approach to improve noise suppression in images.

Keywords

image processing, noise suppression, Markov Random Fields, applied probability, Hőlder regularity, random number, Generalized Cross Validation

Authors

ČERVINKA, T.

RIV year

2004

Released

1. 1. 2004

Location

Praha

Pages from

20

Pages to

25

Pages count

6

URL

http://radio.feld.cvut.cz/conf/poster2004/

BibTex

@inproceedings{BUT13290,
  author="Tomáš {Červinka}",
  title="Generalized Cross Validation with Bayes Approach for Image Denoising",
  booktitle="8th International Student Conference on Electrical Engineering",
  year="2004",
  number="1",
  pages="6",
  address="Praha",
  url="http://radio.feld.cvut.cz/conf/poster2004/"
}