IVML  
  about | r&d | publications | courses | people | links
   

E. Douglas-Cowie, R. Cowie, I. Sneddon, C. Cox, O. Lowry, M. McRorie, J.-C. Martin, L. Devillers, S. Abrilian, A. Batliner, N. Amir, K. Karpouzis
The HUMAINE Database: Addressing the Collection and Annotation of Naturalistic and Induced Emotional Data
Proc. 2nd International Conference on Affective Computing and Intelligent Interaction (ACII'2007), Lisbon, Portugal.
ABSTRACT
A key goal of the HUMAINE project was to provide the community with examples of the diverse data types that are potentially relevant to affective computing, and the kinds of labelling scheme that address the data. Data has been collected to show emotion in a range of contexts. The database proper is a selected subset of the data with systematic labelling, mounted on the ANVIL platform. It is designed to provide a concrete illustration of key principles rather than to be used as it stands in machine learning. Stage 1 (available via the HUMAINE portal at www.emotion-research.net) contains 50 Ąclipsą from naturalistic and induced data, showing a range of modalities and emotions, and covering a balanced sample of emotional behaviour in a range of contexts. Emotional content is described by a structured set of labels attached to the clips both at a global level, and frame-by-frame, showing change over time. Labels for a range of signs of emotion have also been developed and applied to a subset of the clips: these include core signs in speech and language, and descriptors for gestures and facial features that draw on standard descriptive schemes.
19 September, 2007
E. Douglas-Cowie, R. Cowie, I. Sneddon, C. Cox, O. Lowry, M. McRorie, J.-C. Martin, L. Devillers, S. Abrilian, A. Batliner, N. Amir, K. Karpouzis, "The HUMAINE Database: Addressing the Collection and Annotation of Naturalistic and Induced Emotional Data", Proc. 2nd International Conference on Affective Computing and Intelligent Interaction (ACII'2007), Lisbon, Portugal.
[ save PDF] [ BibTex] [ Print] [ Back]

© 00 The Image, Video and Multimedia Systems Laboratory - v1.12