Earticle

초록

영어

This paper proposes a new feature vector selection method for voice pattern recognition tasks, especially for speaker or emotion recognition. During the model training phase, robust speaker or emotion models are constructed by using meaningful feature vectors while discarding confusing vectors that may induce recognition error. To select meaningful feature vectors, the proposed method classifies feature vectors into overlapped and non-overlapped sets using log-likelihood ratio. Speaker- and emotion-recognition experiments confirmed that these robust models significantly reduce recognition errors.

목차

Abstract
 1. Introduction
 2. The Conventional GMM-based Pattern Recognition Technique
  2.1. Principle of GMM-based Speaker and Emotion Recognition
  2.2. Feature Vector Selection
 3. Feature Vector Selection based on Log-Likelihood
 4. Experiments and Results
  4.1. Speaker Recognition Experiments
  4.2. Emotion Recognition Experiments
 5. Conclusions
 References

키워드

저자

• Jeong-Sik Park [ Department of Information and Communication Engineering, Yeungnam University, Republic of Korea ]
• Gil-Jin Jang [ School of Electronics Engineering, Kyungpook National University, Republic of Korea ] Corresponding Author
• Ji-Hwan Kim [ Department of Computer Science and Engineering, Sogang University, Republic of Korea ]

참고문헌

발행기관

간행물

표지보기 관심저널 등록

• 간행물명

  International Journal of Software Engineering and Its Applications

• 간기

  월간

• pISSN

  1738-9984

• 수록기간

  2008~2016

• 등재여부

  SCOPUS

• 십진분류

  KDC 505 DDC 605