SECOND ORDER VECTOR TAYLOR SERIES BASED ROBUST SPEECH RECOGNITION

被引：0
|
作者：
Bu, Suliang [1 ]
Qian, Yanmin [1 ]
Sim, Khe Chai [2 ]
You, Yongbin [1 ]
Yu, Kai [1 ]
机构：
[1] Shanghai Jiao Tong Univ, Dept Comp Sci & Engn, MOE, Microsoft Key Lab Intelligent Comp & Intelligent, Shanghai 200030, Peoples R China
[2] Natl Univ Singapore, Dept Comp Sci, Singapore 117548, Singapore
来源：
2014 IEEE INTERNATIONAL CONFERENCE ON ACOUSTICS, SPEECH AND SIGNAL PROCESSING (ICASSP) | 2014年
关键词：
robust speech recognition; model based compensation; Vector Taylor Series; APPROXIMATION;
D O I：
暂无
中图分类号：
O42 [声学];
学科分类号：
070206 ; 082403 ;
摘要：
Vector Taylor Series (VTS) model based compensation approach has been successfully applied to various robust speech recognition tasks. In this paper, a novel method to derive the formula to calculate the static and dynamic statistics based on second-order VTS (sVTS) is presented, which provides a new insight on the VTS approximation. Lengthy derivation could therefore be avoided when high order VTS is used and the proposed approach is more compact and easier to implement compared to previous high order VTS approaches. Experiments on Aurora 4 showed that the proposed sVTS based model compensation approach obtained 16.7% relative WER reduction over traditional first-order VTS (fVTS) approach.
引用
收藏
页数：5
相关论文
共 50 条
  • [21] A Feature Compensation Approach Using High-Order Vector Taylor Series Approximation of an Explicit Distortion Model for Noisy Speech Recognition
    Du, Jun
    Huo, Qiang
    INTERSPEECH 2008: 9TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION 2008, VOLS 1-5, 2008, : 1257 - 1260
  • [22] Cepstral vector normalization based on stereo data for robust speech recognition
    Buera, Luis
    Lleida, Eduardo
    Miguel, Antonio
    Ortega, Alfonso
    Saz, Oscar
    IEEE TRANSACTIONS ON AUDIO SPEECH AND LANGUAGE PROCESSING, 2007, 15 (03): : 1098 - 1113
  • [23] Using vector Taylor series with noise clustering for speech recognition in non-stationary noisy environments
    Department of Electronic Engineering, Tsinghua University, Beijing 100084, China
    High Technol Letters, 2006, 1 (18-23):
  • [24] Trellis encoded vector quantization for robust speech recognition
    Chou, W
    Seshadri, N
    Rahim, M
    ICSLP 96 - FOURTH INTERNATIONAL CONFERENCE ON SPOKEN LANGUAGE PROCESSING, PROCEEDINGS, VOLS 1-4, 1996, : 2001 - 2004
  • [25] Detecting and Labeling Speakers on Overlapping Speech using Vector Taylor Series
    Dighe, Pranay
    Ferras, Marc
    Bourlard, Herve
    15TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2014), VOLS 1-4, 2014, : 592 - 596
  • [26] Taylor Series Method for Second-Order Polynomial ODEs
    Latypov, Viktor
    Sokolov, Sergei
    2015 INTERNATIONAL CONFERENCE "STABILITY AND CONTROL PROCESSES" IN MEMORY OF V.I. ZUBOV (SCP), 2015, : 62 - 64
  • [27] Emotion recognition of speech signal using Taylor series and deep belief network based classification
    Valiyavalappil Haridas, Arul
    Marimuthu, Ramalatha
    Sivakumar, V. G.
    Chakraborty, Basabi
    EVOLUTIONARY INTELLIGENCE, 2022, 15 (02) : 1145 - 1158
  • [28] Channel Compensation in the Generalised Vector Taylor Series Approach to Robust ASR
    Loweimi, Erfan
    Barker, Jon
    Hain, Thomas
    18TH ANNUAL CONFERENCE OF THE INTERNATIONAL SPEECH COMMUNICATION ASSOCIATION (INTERSPEECH 2017), VOLS 1-6: SITUATED INTERACTION, 2017, : 2466 - 2470
  • [29] Robust speech recognition based on independent vector analysis using harmonic frequency dependency
    Jun, Soram
    Kim, Minook
    Oh, Myungwoo
    Park, Hyung-Min
    NEURAL COMPUTING & APPLICATIONS, 2013, 22 (7-8): : 1321 - 1327
  • [30] Robust speech recognition based on independent vector analysis using harmonic frequency dependency
    Soram Jun
    Minook Kim
    Myungwoo Oh
    Hyung-Min Park
    Neural Computing and Applications, 2013, 22 : 1321 - 1327
12345