Přístupnostní navigace
E-application
Search Search Close
Publication detail
RAJMIC, P.
Original Title
Využití waveletové transformace a matematické statistiky pro separaci signálu a šumu
English Title
Utilization of wavelet transform and mathematical statistics for separating signal and noise
Type
miscellaneous
Language
Czech
English abstract
The dissertation thesis considers a method for suppressing Gaussian (normally distributed) noise in signals utilizing the so-called thresholding of wavelet coefficients, with particular focus on speech signals. This type of separation of effective and noisy signals starts from the WaveShrink methodology. The first part of the thesis considers several modifications of this methodology in order to improve its performance, and, also, in particular the statistical analysis and comparison of five thresholding rules. The second part is devoted to the newly-developed method of so-called segmented wavelet transform, which allows us to perform the wavelet transform, and thus also the "wavelet-type separation" in real time, in addition to other benefits.
Keywords
waveletová transformace, SegWT, denoising
Key words in English
wavelet transform, SegWT, denoising
Authors
Released
1. 1. 2005
Publisher
Vysoké učení technické v Brně
Location
Brno
ISBN
80-214-2810-4
Edition
Edice PhD Thesis, sv. 289
1213-4198
Periodical
Vědecké spisy Vysokého učení technického v Brně Edice PhD Thesis
State
Czech Republic
Pages from
1
Pages to
30
Pages count
URL
Knihovna FEKT VUT v Brně
BibTex
@misc{BUT67129, author="Pavel {Rajmic}", title="Využití waveletové transformace a matematické statistiky pro separaci signálu a šumu", year="2005", series="Edice PhD Thesis, sv. 289", journal="Vědecké spisy Vysokého učení technického v Brně Edice PhD Thesis", pages="30", publisher="Vysoké učení technické v Brně", address="Brno", isbn="80-214-2810-4", issn="1213-4198", url="Knihovna FEKT VUT v Brně", note="miscellaneous" }