The Journal of the Acoustical Society of Korea (한국음향학회지)

The Acoustical Society of Korea (한국음향학회)
권호 표지

State-Dependent Weighting of Multiple Feature Parameters in HMM Recognizer

HMM 인식기에서 상태별 다중 특징 파라미터 가중

  • 손종목 (경북대학교 전자공학과) ;
  • 배건성 (경북대학교 전자공학과)
  • Published : 1999.05.01
Download PDF
⟨ Previous Next ⟩

Abstract

In this paper, we proposed a new approach to weight each feature parameter by considering the dispersion of feature parameters and its degree of contribution to recognition rate. We determined the total distribution factor that is proportional to recognition rate of each feature parameter and the dispersion factor according to the dispersion of each feature parameter. Then. we determined state-dependent weighting using the total distribution factor and dispersion factor. To verify the validity of the proposed approach, recognition experiments were performed using the PLU(Phoneme-Like Unit)-based HMM. Experimental results showed the improvement of 7.7% at the recognition rate using the proposed method.

본 논문에서는 특징 파라미터의 분산과 인식성능에 대한 기여도를 고려하여 각 특징 파라미터를 가중시키는 방법을 제안하였다. 각 특징 파라미터의 인식률에 비례하게 전체 기여도를 설정하고, 각 특징 파라미터의 분산에 따라 가중요인을 설정하였다. 전체 기여도와 분산에 따른 가중요인을 사용하여 각 특징 파라미터의 상태별 가중치를 설정하였다. 제안한 방법의 유효성을 살펴보기 위해 유사음소 단위의 HMM 음성인식시스템을 사용하여 인식실험을 하였다. 인식실험에서 제안한 방법으로 가중치를 설정하였을 경우에 인식률이 7.7% 향상됨을 볼 수 있었다.

Keywords

References

  1. EUROSPEECH v.1 Codebook Wrights Adaptation for Discriminative Training of SCHMM-Based Speech Recognition Systems C.M.del Alamo;F.J.Caminero-Gil;C.dela Torne-Munilla;L.Hernandez-Gomez
  2. ICASSP v.1 Learning State-Dependent Stream Weights for Multi-Codebook HMM Speech Recognition System I.Rogina;A.Waibel
  3. EUROSPEECH v.1 Optimization of Speech Parameter Weighting for CDHMM Word Recognition J.Hernando;J.Ayarte;E.Monte
  4. ICASSP v.2 Maximum Likelihood Weighting of Dynamic Speech Features for CDHMM Speech Recognition J.Hernando
  5. Proc. of IEEE v.77 no.2 A Tutorial on Hidden Markov Models and Selected Applications in Speech Recognition L.R.Rabiner
  6. IEEE Trans. on ASSP v.38 no.9 The Segmental K-Means Algorithm for Estimation Parameters of Hidden Markov Models B.H.Juang;L.R.Rabiner
  7. IEEE Trans. ASSP v.38 no.1 An Overview of the SPHINX Speech Recognition System K.F.Lee;H.W.Hon;R.Reddy
  8. IEEE Trans. on SAP v.1 no.3 A Speaker-Independent Continuous Speech Recognition System Using Continuous Mixture Gaussian Density HMM of Phoneme-Sized Units Y.Zhao
  9. ICASSP v.2 Acoustic Modeling of Subword Units for Speech Recognition C.H.Lee;L.R.Rabiner;R.Pieraccinit;Jay G.Wilpon