Understanding of speakers’ attitude is essential for establishing successful human interaction. In this paper we analyse attitude manifestations in video blogs. We describe the main features of this novel communication medium and focus our attention on its possible exploitation as a rich source of information for human-human and human-machine communication. We describe the manual annotation of attitudes and the prosodic analyses. Finally we present a preliminary attitude automatic annotation system that attains 65% accuracy.
Cite as: Madzlan, N.A., Han, J., Bonin, F., Campbell, N. (2014) Towards Automatic Recognition of Attitudes: Prosodic Analysis of Video Blogs. Proc. Speech Prosody 2014, 91-94, doi: 10.21437/SpeechProsody.2014-6
@inproceedings{madzlan14_speechprosody,
author={Noor Alhusna Madzlan and Jingguang Han and Francesca Bonin and Nick Campbell},
title={{Towards Automatic Recognition of Attitudes: Prosodic Analysis of Video Blogs}},
year=2014,
booktitle={Proc. Speech Prosody 2014},
pages={91--94},
doi={10.21437/SpeechProsody.2014-6}
}