An improvement in audio-visual voice activity detection for automatic speech recognition

Takami Yoshida; Kazuhiro Nakadai; Hiroshi G. Okuno

doi:10.1007/978-3-642-13022-9_6

An improvement in audio-visual voice activity detection for automatic speech recognition

Takami Yoshida^*, Kazuhiro Nakadai, Hiroshi G. Okuno

^*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

6 Citations (Scopus)

Abstract

Noise-robust Automatic Speech Recognition (ASR) is essential for robots which are expected to communicate with humans in a daily environment. In such an environment, Voice Activity Detection (VAD) strongly affects the performance of ASR because there are many acoustically and visually noises. In this paper, we improved Audio-Visual VAD for our two-layered audio visual integration framework for ASR by using hangover processing based on erosion and dilation. We implemented proposed method to our audio-visual speech recognition system for robot. Empirical results show the effectiveness of our proposed method in terms of VAD.

Original language	English
Title of host publication	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Pages	51-61
Number of pages	11
Volume	6096 LNAI
Edition	PART 1
DOIs	https://doi.org/10.1007/978-3-642-13022-9_6
Publication status	Published - 2010
Externally published	Yes
Event	23rd International Conference on Industrial Engineering and Other Applications of Applied Intelligence Systems, IEA/AIE 2010 - Cordoba Duration: 2010 Jun 1 → 2010 Jun 4

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Number	PART 1
Volume	6096 LNAI
ISSN (Print)	03029743
ISSN (Electronic)	16113349

Other

Other	23rd International Conference on Industrial Engineering and Other Applications of Applied Intelligence Systems, IEA/AIE 2010
City	Cordoba
Period	10/6/1 → 10/6/4

Keywords

Audio-Visual integration
Speech Recognition
Voice Activity Detection

ASJC Scopus subject areas

Computer Science(all)
Theoretical Computer Science

Access to Document

10.1007/978-3-642-13022-9_6

Cite this

Yoshida, T., Nakadai, K., & Okuno, H. G. (2010). An improvement in audio-visual voice activity detection for automatic speech recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (PART 1 ed., Vol. 6096 LNAI, pp. 51-61). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6096 LNAI, No. PART 1). https://doi.org/10.1007/978-3-642-13022-9_6

An improvement in audio-visual voice activity detection for automatic speech recognition. / Yoshida, Takami; Nakadai, Kazuhiro; Okuno, Hiroshi G.

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6096 LNAI PART 1. ed. 2010. p. 51-61 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 6096 LNAI, No. PART 1).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Yoshida, T, Nakadai, K & Okuno, HG 2010, An improvement in audio-visual voice activity detection for automatic speech recognition. in Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 edn, vol. 6096 LNAI, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), no. PART 1, vol. 6096 LNAI, pp. 51-61, 23rd International Conference on Industrial Engineering and Other Applications of Applied Intelligence Systems, IEA/AIE 2010, Cordoba, 10/6/1. https://doi.org/10.1007/978-3-642-13022-9_6

Yoshida T, Nakadai K, Okuno HG. An improvement in audio-visual voice activity detection for automatic speech recognition. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). PART 1 ed. Vol. 6096 LNAI. 2010. p. 51-61. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1). doi: 10.1007/978-3-642-13022-9_6

Yoshida, Takami ; Nakadai, Kazuhiro ; Okuno, Hiroshi G. / An improvement in audio-visual voice activity detection for automatic speech recognition. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 6096 LNAI PART 1. ed. 2010. pp. 51-61 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 1).

@inproceedings{2c8917c0f4e0406a9df5b52c9227a020,

title = "An improvement in audio-visual voice activity detection for automatic speech recognition",

abstract = "Noise-robust Automatic Speech Recognition (ASR) is essential for robots which are expected to communicate with humans in a daily environment. In such an environment, Voice Activity Detection (VAD) strongly affects the performance of ASR because there are many acoustically and visually noises. In this paper, we improved Audio-Visual VAD for our two-layered audio visual integration framework for ASR by using hangover processing based on erosion and dilation. We implemented proposed method to our audio-visual speech recognition system for robot. Empirical results show the effectiveness of our proposed method in terms of VAD.",

keywords = "Audio-Visual integration, Speech Recognition, Voice Activity Detection",

author = "Takami Yoshida and Kazuhiro Nakadai and Okuno, {Hiroshi G.}",

year = "2010",

doi = "10.1007/978-3-642-13022-9_6",

language = "English",

isbn = "3642130216",

volume = "6096 LNAI",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

number = "PART 1",

pages = "51--61",

booktitle = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

edition = "PART 1",

note = "23rd International Conference on Industrial Engineering and Other Applications of Applied Intelligence Systems, IEA/AIE 2010 ; Conference date: 01-06-2010 Through 04-06-2010",

}

TY - GEN

T1 - An improvement in audio-visual voice activity detection for automatic speech recognition

AU - Yoshida, Takami

AU - Nakadai, Kazuhiro

AU - Okuno, Hiroshi G.

PY - 2010

Y1 - 2010

N2 - Noise-robust Automatic Speech Recognition (ASR) is essential for robots which are expected to communicate with humans in a daily environment. In such an environment, Voice Activity Detection (VAD) strongly affects the performance of ASR because there are many acoustically and visually noises. In this paper, we improved Audio-Visual VAD for our two-layered audio visual integration framework for ASR by using hangover processing based on erosion and dilation. We implemented proposed method to our audio-visual speech recognition system for robot. Empirical results show the effectiveness of our proposed method in terms of VAD.

AB - Noise-robust Automatic Speech Recognition (ASR) is essential for robots which are expected to communicate with humans in a daily environment. In such an environment, Voice Activity Detection (VAD) strongly affects the performance of ASR because there are many acoustically and visually noises. In this paper, we improved Audio-Visual VAD for our two-layered audio visual integration framework for ASR by using hangover processing based on erosion and dilation. We implemented proposed method to our audio-visual speech recognition system for robot. Empirical results show the effectiveness of our proposed method in terms of VAD.

KW - Audio-Visual integration

KW - Speech Recognition

KW - Voice Activity Detection

UR - http://www.scopus.com/inward/record.url?scp=79551526836&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=79551526836&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-13022-9_6

DO - 10.1007/978-3-642-13022-9_6

M3 - Conference contribution

AN - SCOPUS:79551526836

SN - 3642130216

SN - 9783642130212

VL - 6096 LNAI

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 51

EP - 61

BT - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

T2 - 23rd International Conference on Industrial Engineering and Other Applications of Applied Intelligence Systems, IEA/AIE 2010

Y2 - 1 June 2010 through 4 June 2010

ER -

An improvement in audio-visual voice activity detection for automatic speech recognition

Abstract

Publication series

Other

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this