Für eine korrekte Darstellung dieser Seite benötigen Sie einen XHTML-standardkonformen Browser, der die Darstellung von CSS-Dateien zulässt.


Applications of deep learning to emotion recognition

Frederic Li, University of Siegen 

Date: 19/04/2017

Time: 11:00

Room: H-F 010

A lot of progress has been made for the last few years regarding pattern recognition applications, mainly due to the increasingly powerful hardware and computational power, which allowed promising and powerful pattern recognition models to rise, such as Deep Neural Networks (DNN). But in spite of those advances, the field of emotion recognition still remains largely unexplored. The lack of an established theoretical framework to rigorously define emotions, as well as the scarcity and difficulty of acquisition of the emotion-related data are some of the main obstacles hindering the application of DNN to emotion recognition problems, and preventing the obtention of good performances. But some methods based on unsupervised and transfer learning could be employed bypass this issue, and hopefully obtain better classification results for emotions. This presentation will provide a short overview of the current research field, as well as the reasoning behind the application of DNN based solutions to emotion recognition.