Ego-Centric Emotion Recognition using Augmented Reality Headsets
Abstract
The project will generate a multimodal fused recognition system for live continuous estimate of user affect of augmented reality (AR) head?mounted display (HMD) users. This work will focus on the estimation of a user s valance and arousal which can then be used to estimate higher?level emotions. The data?streams used will be egocentric video, gesture, speech, head position, movement velocity, and other sensor information found on current consumer?available AR?HMDs. The primary outcome of this work will be a live continuous emotion recognizer that will have a demo application available for use on the Microsoft HoloLens 2 optical see?through AR?HMD.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Nov 05, 2021
- Source ID
- HR00112110012
Entities
People
- Yejin Choi
Organizations
- Defense Advanced Research Projects Agency
- University of Washington