LaurenceDevillers's Home Page

Laurence Devillers' research activities have been focused on several aspect of speech processing (acoustic, phonetic, linguistic, ect.) including all the components of a spoken dialog system (speech recognition and understanding, dialogs strategies and the design of spoken interactive systems). Her current research interests include analysis and detection of emotional behaviours in spontaneous speech corpus of real Human-Human dialogs and in audiovisual data.

Key research interests:
- Study of Real-life emotions in audio and multimodal contexts (call-center, TV clips)
- Complex Emotions Representation (Blended, Masked, Conflictual, etc.)
- Multi-level Annotation Scheme: Emotion, Context and Multimodal cues
- Detection of Emotion in Spoken Interaction using Prosodic, Lexical and Non-verbal cues
- Using annotations of corpus as one source of knowledge for the specification of multimodal emotional behavior in ECA.