國立高雄大學圖資館 |

Language: English

Back

經驗模態分解法應用在情緒語音特徵值之計算 = Applications ...

國立高雄大學資訊工程學系碩士班

經驗模態分解法應用在情緒語音特徵值之計算 = Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features

Record Type:	Language materials, printed : monographic
Paralel Title:	Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features
Author:	李英瑋,
Secondary Intellectual Responsibility:	國立高雄大學
Place of Publication:	[高雄市]
Published:	撰者;
Year of Publication:	2012[民101]
Description:	81面部份彩圖，表格 : 30公分;
Subject:	隱藏式馬可夫模型
Subject:	Hidden Markov Model
Online resource:	http://handle.ncl.edu.tw/11296/ndltd/20667204182468877794
Notes:	參考書目：面68-73
Summary:	本論文結合經驗模態分解法(Empirical Mode Decomposition, EMD)與梅爾倒頻譜參數(Mel-Scale Frequency Cepstral Coefficients, MFCC)計算情緒語音特徵值，改善情緒語音之辨識率。EMD將情緒語音訊號分解成多個本質模態函數(Intrinsic Mode Function, IMF)，並且以三種演化式計算(Evolutional Computation, EC)演算法分別為粒子群演算法(Particle Swarm Optimization, PSO)、基因演算法(Genetic Algorithm, GA)以及差分演算法(Differential Evolution, DE)，計算出每個IMF之最佳權重值組合，以強化情緒語音訊號。另外，我們實驗使用隱藏式馬可夫模型(Hidden Markov Model, HMM)訓練以及辨識情緒語音特徵值。由實驗結果得知，本論文所提出之方法的確可以改善情緒語音之辨識率。 This thesis combines Empirical Mode Decomposition (EMD) with Mel-Scale Frequency Cepstral Coefficients (MFCC) to extract emotional speech features and improve emotional speech recognition rate. The EMD method is used to decompose emotional speech signals into several Intrinsic Mode Functions (IMFs). Three evolutionary algorithms: Particle Swarm Optimization (PSO), Differential Evolution (DE), and Genetic Algorithm (GA) are used to find the optimal weights of IMFs to compose an enhanced emotional speech signal. Thereafter, we can obtain more suitable emotional features by using MFCC. After extracting features, we fed these features into the Hidden Markov Model (HMM) for training and testing. Finally, experimental results will show that the emotional speech recognition rate can be improved by using the proposed method.

經驗模態分解法應用在情緒語音特徵值之計算 = Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features
李, 英瑋

經驗模態分解法應用在情緒語音特徵值之計算 = Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features / 李英瑋撰 - [高雄市] : 撰者, 2012[民101]. - 81面 ; 部份彩圖，表格 ; 30公分.
參考書目：面68-73.
隱藏式馬可夫模型Hidden Markov Model

經驗模態分解法應用在情緒語音特徵值之計算 = Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features
LDR:02960nam0a2200277 450 001 346197
005 20170214095950.0
009 346197
010 0 $b 精裝
010 0 $b 平裝
100 $a 20170214d2012 k y0chiy05 e
101 1 $a chi $d chi $d eng
102 $a tw
105 $a ak am 000yy
200 1 $a 經驗模態分解法應用在情緒語音特徵值之計算 $d Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features $z eng $f 李英瑋撰
210 $a [高雄市] $c 撰者 $d 2012[民101]
215 0 $a 81面 $c 部份彩圖，表格 $d 30公分
300 $a 參考書目：面68-73
314 $a 指導教授：潘欣泰博士
328 $a 碩士論文--國立高雄大學資訊工程學系碩士班
330 $a 本論文結合經驗模態分解法(Empirical Mode Decomposition, EMD)與梅爾倒頻譜參數(Mel-Scale Frequency Cepstral Coefficients, MFCC)計算情緒語音特徵值，改善情緒語音之辨識率。EMD將情緒語音訊號分解成多個本質模態函數(Intrinsic Mode Function, IMF)，並且以三種演化式計算(Evolutional Computation, EC)演算法分別為粒子群演算法(Particle Swarm Optimization, PSO)、基因演算法(Genetic Algorithm, GA)以及差分演算法(Differential Evolution, DE)，計算出每個IMF之最佳權重值組合，以強化情緒語音訊號。另外，我們實驗使用隱藏式馬可夫模型(Hidden Markov Model, HMM)訓練以及辨識情緒語音特徵值。由實驗結果得知，本論文所提出之方法的確可以改善情緒語音之辨識率。 This thesis combines Empirical Mode Decomposition (EMD) with Mel-Scale Frequency Cepstral Coefficients (MFCC) to extract emotional speech features and improve emotional speech recognition rate. The EMD method is used to decompose emotional speech signals into several Intrinsic Mode Functions (IMFs). Three evolutionary algorithms: Particle Swarm Optimization (PSO), Differential Evolution (DE), and Genetic Algorithm (GA) are used to find the optimal weights of IMFs to compose an enhanced emotional speech signal. Thereafter, we can obtain more suitable emotional features by using MFCC. After extracting features, we fed these features into the Hidden Markov Model (HMM) for training and testing. Finally, experimental results will show that the emotional speech recognition rate can be improved by using the proposed method.
510 1 $a Applications of Empirical Mode Decomposition on the Computation of Emotional Speech Features $z eng
610 0 $a 隱藏式馬可夫模型 $a 經驗模態分解法 $a 粒子群演算法 $a 基因演算法 $a 差分演算法 $a 情緒語音辨識
610 1 $a Hidden Markov Model $a Empirical Mode Decomposition $a Particle Swarm Optimization $a Differential Evolution $a Genetic Algorithm $a Emotional Speech Recognition
681 $a 008M/0019 $b 464103 4041 $v 2007年版
700 1 $a 李 $b 英瑋 $4 撰 $3 576411
712 0 2 $a 國立高雄大學 $b 資訊工程學系碩士班 $3 353878
801 0 $a tw $b NUK $c 20121107 $g CCR
856 7 $z 電子資源 $2 http $u http://handle.ncl.edu.tw/11296/ndltd/20667204182468877794