Neural Graph Matching Networks for Fewshot 3D Action Recognition
Neural Graph Matching Networks for Fewshot 3D Action Recognition
Michelle Guo,Edward Chou,3 Authors,Li Fei-Fei
2018 · DOI: 10.1007/978-3-030-01246-5_40
European Conference on Computer Vision · 109 Citations
TLDR
This work proposes Neural Graph Matching Networks, a novel framework that can learn to recognize a previous unseen 3D action class with only a few examples by leveraging the inherent structure of 3D data through a graphical representation to directly optimize the few-shot learning objective.