+ All Categories
Home > Documents > Tone Recognition With Fractionized Models and Outlined Features

Tone Recognition With Fractionized Models and Outlined Features

Date post: 22-Feb-2016
Category:
Upload: paige
View: 23 times
Download: 0 times
Share this document with a friend
Description:
Tone Recognition With Fractionized Models and Outlined Features. Ye Tian , Jian -Lai Zhou, Min Chu, Eric Chang ICASSP 2004. Hsiao- Tsung Hung. Department of Computer Science and Information Engineering National Taiwan Normal University. Outline. Introduction Features Detailed features - PowerPoint PPT Presentation
Popular Tags:
16
Tone Recognition With Fractionized Models and Outlined Features Ye Tian, Jian-Lai Zhou, Min Chu, Eric Chang ICASSP 2004 Hsiao-Tsung Hung Department of Computer Science and Information Engineering National Taiwan Normal University
Transcript
Page 1: Tone Recognition With Fractionized Models and Outlined Features

Tone Recognition With Fractionized Models and Outlined Features

Ye Tian, Jian-Lai Zhou, Min Chu, Eric ChangICASSP 2004

Hsiao-Tsung Hung

Department of Computer Science and Information EngineeringNational Taiwan Normal University

Page 2: Tone Recognition With Fractionized Models and Outlined Features

Outline

• Introduction• Features– Detailed features– Outlined features– Experiments and analysis

• Tone Modeling– Experiments and analysis

• Conclusions

Page 3: Tone Recognition With Fractionized Models and Outlined Features

Introduction

• 2 questions1. Is the detailed information of F0 curve useful

for tone discrimination in continuous speech?

2. Are phoneme-independent tone models sufficient for continuous speech recognition?

Page 4: Tone Recognition With Fractionized Models and Outlined Features

Detailed features

• Detailed features: Using the entire F0 curve.• Observation vector is • If the phoneme has totally N frames, the

number of total parameters used for tone recognition is 2*N.

Page 5: Tone Recognition With Fractionized Models and Outlined Features

Outlined features

• To reduce the number of parameters and improve the robustness.

1. Curve fitting features2. Subsection Outlined features

Page 6: Tone Recognition With Fractionized Models and Outlined Features

Curve fitting features

• First-order

• Second-order

Page 7: Tone Recognition With Fractionized Models and Outlined Features

Subsection Outlined features

• The F0 curve of the entire phoneme is divided into several subsections and each subsection is represented by certain parameters.

• Extract parameters for each subsection– 1.subsection slop and intercept– 2.subsection and (Assume that time frames belong to the subsection k.)

Page 8: Tone Recognition With Fractionized Models and Outlined Features

Y

X

X={0,1,…, //frame

Y=

F0

Page 9: Tone Recognition With Fractionized Models and Outlined Features

Subsection Outlined features

1.subsection slop and intercept

Page 10: Tone Recognition With Fractionized Models and Outlined Features

Subsection Outlined features

2.subsection and

Page 11: Tone Recognition With Fractionized Models and Outlined Features
Page 12: Tone Recognition With Fractionized Models and Outlined Features
Page 13: Tone Recognition With Fractionized Models and Outlined Features

Experiments and analysis

1.Main value and direction are the most important characteristics.

2.Detailed information is useless for tone discrimination.

Page 14: Tone Recognition With Fractionized Models and Outlined Features

Tone Modeling

1. One-tone-one-model tone models(5)2. Monophone-dependent tone models(54)

The same tone in different tonal phonemes is different modeled.

3. Triphone-dependent tone models(12824)

Page 15: Tone Recognition With Fractionized Models and Outlined Features

Experiments and analysis

• Feature vector :

Page 16: Tone Recognition With Fractionized Models and Outlined Features

Conclusions

• Using fractionized models and outlined features for tone recognition.

• Outlined features can reduce the interference caused by co-articulation effect, syllable stress, and sentence intonation.


Recommended