Generative Range and Altitude Identity Learning

Idiap investigate integrated fusion approaches to Deep Neural Network (DNN)-based whole-body biometrics combining gait and face recognition. We explore fusion strategies to combine face and whole-body representations integrating ancillary information from individual modalities to cope with missing channels and adopt contrastive training (eg. cross-modal focal loss) to factor the confidence of other modalities.
University of Southern California
Idiap Research Institute
IARPA
May 12, 2023
Jul 04, 2024