Open Access. Powered by Scholars. Published by Universities.®
Articles 1 - 1 of 1
Full-Text Articles in Entire DC Network
Multimodal Emotion Recognition Using 3d Facial Landmarks, Action Units, And Physiological Data, Diego Fabiano
Multimodal Emotion Recognition Using 3d Facial Landmarks, Action Units, And Physiological Data, Diego Fabiano
USF Tampa Graduate Theses and Dissertations
To fully understand the complexities of human emotion, the integration of multiple physical features from different modalities can be advantageous. Considering this, this thesis presents an approach to emotion recognition using handcrafted features that consist of 3D facial data, action units, and physiological data. Each modality independently, as well as the combination of each for recognizing human emotion were analyzed.
This analysis includes the use of principal component analysis to determine which dimensions of the feature vector are most important for emotion recognition. The proposed features are shown to be able to be used to accurately recognize emotion and that …