확률출력 SVM을 이용한 감정식별 및 감정검출

Identification and Detection of Emotion Using Probabilistic Output SVM

  • 조훈영 (한국전자통신연구원 디지털콘텐츠연구단 HD게임연구팀) ;
  • 정규준 (한국과학기술원 전자전산학과)
  • 발행 : 2006.11.30

초록

본 논문에서는 음성신호에 포함된 감정정보를 자동으로 식별하는 방법과 특정 감정을 검출하는 방법에 대해 다룬다. 자동 감정식별 및 검출을 위해 장구간 (long-term) 음향 특징을 사용하였고, F-score 기반의 특징선택 기법을 적용하여 최적의 특징 파라미터들을 선정하였다. 기존의 일반적인 SVM을 확률출력 SVM으로 변환하여 감정식별 및 감정검출 시스템을 구축하였으며, 가설검정에 기반한 감정검출을 위해 세 가지의 대수 우도비 (log-likelihood) 근사법을 제안하여 그 성능을 비교하였다. SUSAS 데이터베이스를 사용한 실험 결과, F-score를 이용한 특징선택 기법에 의해 감정식별 성능이 향상되었으며, 확률출력 SVM의 유효성을 검증할 수 있었다. 감정검출의 경우, 제안한 방법에 의해 91.3%의 정확도로 화난 감정을 검출할 수 있었다.

This paper is about how to identify emotional information and how to detect a specific emotion from speech signals. For emotion identification and detection task. we use long-term acoustic feature parameters and select the optimal Parameters using the feature selection technique based on F-score. We transform the conventional SVM into probabilistic output SVM for our emotion identification and detection system. In this paper we propose three approximation methods for log-likelihoods in a hypothesis test and compare the performance of those three methods. Experimental results using the SUSAS database showed the effectiveness of both feature selection and Probabilistic output SVM in the emotion identification task. The proposed methods could detect anger emotion with 91.3% correctness.

키워드

참고문헌

  1. K. Scherer,'Vocal communication of emotion: A review of research paradigms', Speech Communications, 40 227-256, 2003 https://doi.org/10.1016/S0167-6393(02)00084-5
  2. N. Amir,S. Ziv, and R. Cohen,'Characteristics of authentic anger in Hebrew speech',In Proc. ,Eurospeech, 713-716, 2003
  3. S. Yacoub,S. Skimske,and J. Burns,Recognition of emotions in interactive voice response systems',In Proc Eurospeech,729-732,2003
  4. O.-W. Kwon,K. Chan,J. Hao, and T. W. Lee,'Emotion recognition by speech signals',In Proc. Eurospeech,125-128, 2003
  5. I. Guvon and A. Elisseef,'An introduction to variable and feature selection', Journal of Machine Learning Research,3 1157-1182,2003 https://doi.org/10.1162/153244303322753616
  6. A. Rakotomamonjy,'Variable selection using SVM based criteria',Journal of Machine Learning Research, 3 1357-1370, 2003 https://doi.org/10.1162/153244303322753706
  7. R. A. Sukkar and C. H. Lee, 'Vocabulary independent discriminative utterance verification for non keyword rejection in subword based speech recognition', IEEE Speech and Audio Processing, 4(6) 420-429, 1996 https://doi.org/10.1109/89.544527
  8. J. Platt, 'Probabilistic outputs for support vector machines and comparison to regularized likelihood methods', Advances in Large Margin Classifiers, MIT Press, 2000
  9. D. Price,S. Knerr, L. Personnaz,and G. DreyFus,'Pairwise neural network classifiers with probabilistic outputs', Advances in Neural Information Processing Systems, 1109-1116, 1995
  10. R. Tato,R. Santos,R. Kompe,and J. M. Pardo,'Emotional space improves emotion recognition',In Proc. of International Conference on Spoken Language Processing, 2029-2032, 2002
  11. T. Robinson,'Cookbook: Speech processing functions in C' CD ROM: Prime Time Freeware for AI,Issue 1-1, 1994