上一条： Spatial-temporal graph attention networks for skeleton-based action recognition

下一条： Meta-Learning based prototype-relation network for few-shot classification