| Bookmark: |
Full-text
| URL: http://www.pyoudeyer.com/prosodyRec.ps |
| Cached: PDF-32K PS-63K PS.gz-25K |
| SAVE AS an easy-to-recall long filename: |
| Filename format: author--year--title PDF-32K PS-25K :: About GZip'd PS |
| Filename format: author--year--title--journal|proceedings|...--pages PDF-32K PS-25K |
Related links
| Web search: Google Web Search :: Google Scholar |
| Within this site: References (9) |
Paper at a Glance
BibTexNovel Useful Features and Algorithms for the Recognition of Emotions in Human Speech Pierreyves Oudeyer Sony Computer Science Lab, Paris, France py@csl.sony.frAbstract Recent years have been marked by the development of robotic pets or partners such as small animals or humanoids. People interact with them using natural human social cues, in particu lar emotional expressions. It is crucial that robot can detect the emotional information contained in speech using only prosodic features, since this is often the only information that they can measure. We present here the first large scale experiment in which a large feature set space and a large machine learning al gorithm space are searched concurrently. We describe new fea tures which prove to be much more efficient than the traditional features used in the litterature.
1. Introduction Recent years have been marked by the increasing development of personal robots, either used as new educational technologies or for pure entertainment/ Typically, these robots look like fa miliar pets such as dogs or cats (e.g. the Sony AIBO robot), or sometimes take the shape of young children such as the hu manoids SDR5 (Sony). Among the capabilities that these personal robots need, one of the most basic is the ability recognize human emotions. In deed, not only emotions are crucial to human reasoning, but they are central to social regulation. Emotional communication is at the same time primitive enough and efficient enough so that we use it a lot when we interact with pets, in particular when we tame them. This is also certainly what allows children to boot strap language learning and should be inspiring to teach robots natural language. In this paper, we present a set of experiments that formed the basis of a technology for automatically recognizing the emo tions in speech based on prosodic features, and used now in certain entertainment robots such as the Sony AIBO or SDR4. This work is the first (to our knowledge) ...
@inproceedings{oudeyerprosody2002a,
author={P-Y. Oudeyer},
title={Novel Useful Features and Algorithms for the Recognition of Emotions in Speech},
year={2002},
pages={547-550},
editor={Bel B. and Marlien I.},
booktitle={Proceedings of the 1st International Conference on Speech Prosody},
url={http://www.isrl.uiuc.edu/~amag/langev/paper/oudeyerprosody2002a.html}
}
| HOME :: Conference List :: Conference Paper | Comments to: junwang4 you-know-at gmail.com | Last update: 2/6/08 |