SCORE NORMALIZATION AND SYSTEM COMBINATION FOR IMPROVED KEYWORD SPOTTING

被引：0

作者：

Karakos, Damianos ^{[1
]}

Schwartz, Richard ^{[1
]}

Tsakalidis, Stavros ^{[1
]}

Zhang, Le ^{[1
]}

Ranjan, Shivesh ^{[1
]}

Ng, Tim ^{[1
]}

Hsiao, Roger ^{[1
]}

Saikumar, Guruprasad ^{[1
]}

Bulyko, Ivan ^{[1
]}

Long Nguyen ^{[1
]}

Makhoul, John ^{[1
]}

Grezl, Frantisek ^{[2
]}

Hannemann, Mirko ^{[2
]}

Karafiat, Martin ^{[2
]}

Szoke, Igor ^{[2
]}

Vesely, Karel ^{[2
]}

Lamel, Lori ^{[3
]}

Le, Viet-Bac ^{[4
]}

机构：

[1] Raytheon BBN Technol, Cambridge, MA 02138 USA

[2] Brno Univ Technol, SpeechFIT, Brno, Czech Republic

[3] CNRS LIMSI, Paris, France

[4] Vocapia Res, Paris, France

来源：

2013 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING (ASRU) | 2013年

关键词：

keyword search; score normalization; system combination; indexing and search;

D O I：

暂无

中图分类号：

TP18 [人工智能理论];

学科分类号：

081104 ; 0812 ; 0835 ; 1405 ;

摘要：

We present two techniques that are shown to yield improved Keyword Spotting (KWS) performance when using the ATWV/MTWV performance measures: (i) score normalization, where the scores of different keywords become commensurate with each other and they more closely correspond to the probability of being correct than raw posteriors; and (ii) system combination, where the detections of multiple systems are merged together, and their scores are interpolated with weights which are optimized using MTWV as the maximization criterion. Both score normalization and system combination approaches show that significant gains in ATWV/MTWV can be obtained, sometimes on the order of 8-10 points (absolute), in five different languages. A variant of these methods resulted in the highest performance for the official surprise language evaluation for the IARPA-funded Babel project in April 2013.

引用

页码：210 / 215

页数：6

共 50 条

[41] THE 2013 BBN VIETNAMESE TELEPHONE SPEECH KEYWORD SPOTTING SYSTEM
Tsakalidis, Stavros
Hsiao, Roger
Karakos, Damianos
Ng, Tim
Ranjan, Shivesh
Saikumar, Guruprasad
Zhang, Le
Nguyen, Long
Schwartz, Richard
Makhoul, John
2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP), 2014,
[42] Improving the performance of a keyword spotting system by using support vector
Benayed, Y
Fohr, D
Haton, JP
Chollet, G
ASRU'03: 2003 IEEE WORKSHOP ON AUTOMATIC SPEECH RECOGNITION AND UNDERSTANDING ASRU '03, 2003, : 145 - 149
[43] Improved External Speaker-Robust Keyword Spotting for Hearing Assistive Devices
Lopez-Espejo, Ivan
Tan, Zheng-Hua
Jensen, Jesper
IEEE-ACM TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2020, 28 : 1233 - 1247
[44] Keyword Spotting with Quaternionic ResNet: Application to Spotting in Greek Manuscripts
Sfikas, Giorgos
Retsinas, George
Giotis, Angelos P.
Gatos, Basilis
Nikou, Christophoros
DOCUMENT ANALYSIS SYSTEMS, DAS 2022, 2022, 13237 : 382 - 396
[45] COMBINATION OF SEARCH TECHNIQUES FOR IMPROVED SPOTTING OF OOV KEYWORDS
Karakos, Damianos
Schwartz, Richard M.
2015 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH, AND SIGNAL PROCESSING (ICASSP), 2015, : 5336 - 5340
[46] A multimodel keyword spotting system based on lip movement and speech features
Handa, Anand
Agarwal, Rashi
Kohli, Narendra
MULTIMEDIA TOOLS AND APPLICATIONS, 2020, 79 (27-28) : 20461 - 20481
[47] A depthwise separable convolutional neural network for keyword spotting on an embedded system
Peter Mølgaard Sørensen
Bastian Epp
Tobias May
EURASIP Journal on Audio, Speech, and Music Processing, 2020
[48] A depthwise separable convolutional neural network for keyword spotting on an embedded system
Sorensen, Peter Molgaard
Epp, Bastian
May, Tobias
EURASIP JOURNAL ON AUDIO SPEECH AND MUSIC PROCESSING, 2020, 2020 (01)
[49] A multimodel keyword spotting system based on lip movement and speech features
Anand Handa
Rashi Agarwal
Narendra Kohli
Multimedia Tools and Applications, 2020, 79 : 20461 - 20481
[50] A novel spoken keyword spotting system using support vector machine
Sangeetha, J.
Jothilakshmi, S.
ENGINEERING APPLICATIONS OF ARTIFICIAL INTELLIGENCE, 2014, 36 : 287 - 293

← 1 2 3 4 5 →