Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system

Jhing Fa Wang; Ta Wen Kuan; Jia Ching Wang; Ta Wei Sun

doi:10.1007/978-3-642-13318-3_65

Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system

Jhing Fa Wang, Ta Wen Kuan, Jia Ching Wang, Ta Wei Sun

研究成果: 書貢獻/報告類型 › 會議論文篇章 › 同行評審

8 引文斯高帕斯（Scopus）

摘要

This work proposes a dynamic fixed-point arithmetic design for SVM-based speaker identification in embedded environment. The whole speaker identification system includes LPCC extraction, SVM training with sequential minimal optimization (SMO), and SVM recognition. The proposed dynamic fixed-point design is applied to each arithmetic procedure and fixed-point error analysis is also performed. The fixed-point SVM-based speaker identification system have been implemented and evaluated on ARM9 DMA2400. The experimental results show that the speaker identification accuracy is slightly degraded with the proposed dynamic fixed-point technique.

原文	???core.languages.en_GB???
主出版物標題	Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings
頁面	524-531
頁數	8
版本	PART 2
DOIs	https://doi.org/10.1007/978-3-642-13318-3_65
出版狀態	已出版 - 2010
事件	7th International Symposium on Neural Networks, ISNN 2010 - Shanghai, China 持續時間: 6 6月 2010 → 9 6月 2010

出版系列

名字	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
號碼	PART 2
卷	6064 LNCS
ISSN（列印）	0302-9743
ISSN（電子）	1611-3349

???event.eventtypes.event.conference???

???event.eventtypes.event.conference???	7th International Symposium on Neural Networks, ISNN 2010
國家/地區	China
城市	Shanghai
期間	6/06/10 → 9/06/10

存取文件

10.1007/978-3-642-13318-3_65

引用此

Wang, J. F., Kuan, T. W., Wang, J. C., & Sun, T. W. (2010). Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system. 於 Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings (PART 2 編輯, 頁 524-531). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 6064 LNCS, 編號 PART 2). https://doi.org/10.1007/978-3-642-13318-3_65

Wang, Jhing Fa ; Kuan, Ta Wen ; Wang, Jia Ching 等. / Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system. Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings. PART 2. 編輯 2010. 頁 524-531 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2).

@inproceedings{d9d23077bc1b41a6a86489b29643aaf1,

title = "Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system",

abstract = "This work proposes a dynamic fixed-point arithmetic design for SVM-based speaker identification in embedded environment. The whole speaker identification system includes LPCC extraction, SVM training with sequential minimal optimization (SMO), and SVM recognition. The proposed dynamic fixed-point design is applied to each arithmetic procedure and fixed-point error analysis is also performed. The fixed-point SVM-based speaker identification system have been implemented and evaluated on ARM9 DMA2400. The experimental results show that the speaker identification accuracy is slightly degraded with the proposed dynamic fixed-point technique.",

keywords = "Support vector machine (SVM), dynamic fixed-point design, linear prediction cepstral coefficient (LPCC), sequential minimal optimization (SMO)",

author = "Wang, {Jhing Fa} and Kuan, {Ta Wen} and Wang, {Jia Ching} and Sun, {Ta Wei}",

year = "2010",

doi = "10.1007/978-3-642-13318-3_65",

language = "???core.languages.en_GB???",

isbn = "3642133177",

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

number = "PART 2",

pages = "524--531",

booktitle = "Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings",

edition = "PART 2",

note = "7th International Symposium on Neural Networks, ISNN 2010 ; Conference date: 06-06-2010 Through 09-06-2010",

}

Wang, JF, Kuan, TW, Wang, JC & Sun, TW 2010, Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system. 於 Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings. PART 2 edn, Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 編號 PART 2, 卷 6064 LNCS, 頁 524-531, 7th International Symposium on Neural Networks, ISNN 2010, Shanghai, China, 6/06/10. https://doi.org/10.1007/978-3-642-13318-3_65

Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system. / Wang, Jhing Fa; Kuan, Ta Wen; Wang, Jia Ching 等.
Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings. PART 2. 編輯 2010. p. 524-531 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); 卷 6064 LNCS, 編號 PART 2).

研究成果: 書貢獻/報告類型 › 會議論文篇章 › 同行評審

TY - GEN

T1 - Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system

AU - Wang, Jhing Fa

AU - Kuan, Ta Wen

AU - Wang, Jia Ching

AU - Sun, Ta Wei

PY - 2010

Y1 - 2010

N2 - This work proposes a dynamic fixed-point arithmetic design for SVM-based speaker identification in embedded environment. The whole speaker identification system includes LPCC extraction, SVM training with sequential minimal optimization (SMO), and SVM recognition. The proposed dynamic fixed-point design is applied to each arithmetic procedure and fixed-point error analysis is also performed. The fixed-point SVM-based speaker identification system have been implemented and evaluated on ARM9 DMA2400. The experimental results show that the speaker identification accuracy is slightly degraded with the proposed dynamic fixed-point technique.

AB - This work proposes a dynamic fixed-point arithmetic design for SVM-based speaker identification in embedded environment. The whole speaker identification system includes LPCC extraction, SVM training with sequential minimal optimization (SMO), and SVM recognition. The proposed dynamic fixed-point design is applied to each arithmetic procedure and fixed-point error analysis is also performed. The fixed-point SVM-based speaker identification system have been implemented and evaluated on ARM9 DMA2400. The experimental results show that the speaker identification accuracy is slightly degraded with the proposed dynamic fixed-point technique.

KW - Support vector machine (SVM)

KW - dynamic fixed-point design

KW - linear prediction cepstral coefficient (LPCC)

KW - sequential minimal optimization (SMO)

UR - http://www.scopus.com/inward/record.url?scp=77954413520&partnerID=8YFLogxK

U2 - 10.1007/978-3-642-13318-3_65

DO - 10.1007/978-3-642-13318-3_65

M3 - 會議論文篇章

AN - SCOPUS:77954413520

SN - 3642133177

SN - 9783642133176

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

SP - 524

EP - 531

BT - Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings

T2 - 7th International Symposium on Neural Networks, ISNN 2010

Y2 - 6 June 2010 through 9 June 2010

ER -

Wang JF, Kuan TW, Wang JC, Sun TW. Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system. 於 Advances in Neural Networks - ISNN 2010 - 7th International Symposium on Neural Networks, ISNN 2010, Proceedings. PART 2 編輯 2010. p. 524-531. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); PART 2). doi: 10.1007/978-3-642-13318-3_65

Dynamic fixed-point arithmetic design of embedded SVM-based speaker identification system

摘要

出版系列

???event.eventtypes.event.conference???

存取文件

指紋

引用此