Interspeech 2017 ComPareE

Interspeech is the world’s largest and most comprehensive conference on the science and technology of spoken language processing. Each year, INTERSPEECH hold the Interspeech Computational Paralinguistics Challenge (ComParE) series, which is an open Challenge in the field of Computational Paralinguistics dealing with states and traits of speakers as manifested in their speech signal’s properties.

In 2017, the problem we are going to solve is to automatic detect the speech under the people who got cold. In short, the approach we use is by combining the sub-dictionary features, decision level features and eGeMAPS functional features. Optimize each feature with different classifiers and finally got 70.2% UAR on the development-set.

To see more detail please refer to my blog.


Last Update Apr. 14th 2020

© 2020 by Hui-Ting(Winnie) Hong

All rights reserved.