Využití adaptivního Neuro-Fuzzy inferenčního systému ve zpracování signálů v oblasti zpracování řeči

Abstract

This master’s thesis deals with a design of a software application of adaptive system ANFIS with aim of use in a noise suppresion of speech record. Design and implementation of this adaptive system is described in this work. Efficiency of created application is tested on synthetic and real data with the focus on filtration efficiency under different ANFIS parameters. For objective evaluation of filtration efficiency SNR, SSNR, DTW and RMSE are estimated.

Description

Subject(s)

ANFIS, speech, noise, adaptive system, filtering, LMS, RLS, SNR, SSNR, DTW, RMSE, neural network, C#

Citation