Sixth International Workshop on Models and Analysis of Vocal Emissions for Biomedical Applications (MAVEBA 2009)
In this work we present the design of an Automatic Emotion Recognizer which tries to take advantage of three soft computing techniques: Neural Networks, Fuzzy Inference Systems and Genetic Algorithms in order to indentify and classify emotions from a speech signal. The classification is done between two emotional states: Negative and Idle. The emotion recordings used for this work belongs to the FAU AIBO database, where children interacting with Sony's pet robot Aibo were recorded. We propose and analyze the use of 18 acoustic features. A classification system based on ANFIS is implemented. Genetic Algorithms are used to select features and tune the ANFIS configuration settings. System implementation and some experimental results are shown.
Index Terms. Emotion recognition, ANFIS applications, genetic algorithms, acoustic speech features, feature selection
Full Paper (reprinted with permission from Firenze University Press)
Bibliographic reference. Pérez Espinosa, Humberto / Reyes García, Carlos Alberto (2009): "Detection of negative emotional state in speech with ANFIS and genetic algorithms", In MAVEBA-2009, 25-28.