상세정보
학술대회
Speech Emotion Recognition Framework based on User Self-referential Speech Features
- 발행일
- 201810
- 출처
- Global Conference on Consumer Electronics (GCCE) 2018, pp.341-342
- 협약과제
- 18ZS1100, 자율성장형 AI 핵심원천기술 연구, 이윤근
- 초록
- This paper proposes a SER (Speech Emotion Recognition) framework using user self-referential speech features, which reflect the ROC (Rate of Change) of the corresponding feature value according to the user's emotion states. The experiments shows that using the proposed self-referential features in addition to the absolute value of speech features can improve the SER accuracy.
- KSP 제안 키워드
- Absolute value, Feature values, Self-referential, Speech Emotion recognition, Speech features, rate of change, user's emotion