ISCA Archive Interspeech 2014
ISCA Archive Interspeech 2014

Automatic recognition of attitudes in video blogs — prosodic and visual feature analysis

Noor Alhusna Madzlan, JingGuang Han, Francesca Bonin, Nick Campbell

This paper reports a study of attitude manifestations in video blogs. We describe the manual annotation of speaker attitudes in a corpus of over 130 video blogs and present an analysis of prosodic and visual cues in relation to attitude states. We use machine learning techniques for the automatic prediction of attitudes from prosodic and visual features in video blogs and compare the performance of prosodic and visual feature sets.