The SEMAINE corpus consists of emotionally coloured conversations. Users were recorded while holding conversations with an operator who adopts in sequence four roles designed to evoke emotional reactions. The operator and the user are seated in separate rooms; they see each other through teleprompter screens, and hear each other through speakers. To allow high quality recording, they are recorded by five high-resolution, high framerate cameras, and by four microphones.
Fully continues (i.e. time and value) annotations of 21 dimensions including valence, arousal, power, and the six basic emotion by between three and seven raters per session is included. Full transcription of the conversations is included as well.
The data has been recorded with a great emphasis on synchrony of the various audio and video streams.
The database is about to be extended with another 600 clips. More information can be found in:
Gary McKeown, Michel F. Valstar, Roderick Cowie, and Maja Pantic, ‘The SEMAINE Corpus of Emotionally Coloured Character Interactions’, Proc. IEEE Int’l Conf. Multimedia & Expo (ICME’10), pp. 1079-1084, Singapore, July 2010
- year: 2010
- url: http://semaine-db.eu/
- main_author: SEMAINE
- license: Free for academic purposes
- subjects: 20
- recordings: 140
- duration: 5 minutes
- naturality: natural
- media: Audio-visual: 5 video streams; 4 audio streams
- language: English
- interaction: human-agent
- annotation: 21 dimensional affect, of which 5 dimensions are coded by at least 6 raters, forced-aligned transcripts, some AU, some head nods/shakes, some laughter
Categories: face-analysis; language-analysis; voice-analysis; multimodal-analysis; linguistic-modelling; psychological-modelling