Output details

Recognising realistic emotions and affect in speech: State of the art and lessons learnt from the first challenge

<17>Summarises the first official Challenge on Emotion Recognition from Speech I organised. It laid the foundation for a highly successful series of events (8 by now, 66 teams participating in the current one) and likewise for an invited overview in the highly prestigious IEEE Signal Processing Magazine, an ERC Starting Grant ("iHEARu", 9%/3329, 1.5 Mio EUR, 60 months, in negotiation) to continue driving Computational Paralinguistics, and a book published with Wiley. It also introduces for the first time intelligent data sampling, compares dynamic and static modelling, and shows fusion of engines to reach the best result on the task.