This paper proposes a novel method for early action prediction based on 3D skeleton data. Our method combines the advantages of graph convolutional networks (GCNs) and adversarial learning to avoid the problems of ins...This paper proposes a novel method for early action prediction based on 3D skeleton data. Our method combines the advantages of graph convolutional networks (GCNs) and adversarial learning to avoid the problems of insufficient spatio-temporal feature extraction and difficulty in predicting actions in the early execution stage of actions. In our method, GCNs, which have outstanding performance in the field of action recognition, are used to extract the spatio-temporal features of the skeleton. The model learns how to optimize the feature distribution of partial videos from the features of full videos through adversarial learning. Experiments on two challenging action prediction datasets show that our method performs well on skeleton-based early action prediction. State-of-the-art performance is reported in some observation ratios.展开更多
基金supported by the Natural Science Foundation of Xinjiang Uygur Autonomous Region of China under Grant No.2022D01A59Xinjiang Uygur Autonomous Region University Scientific Research Project(Key Natural Science Project)under Grant No.XJEDU2021I029+1 种基金the National Natural Science Foundation of China under Grant No.U20A20167the Project of Hebei Key Laboratory of Software Engineering under Grant No.22567637H.
文摘This paper proposes a novel method for early action prediction based on 3D skeleton data. Our method combines the advantages of graph convolutional networks (GCNs) and adversarial learning to avoid the problems of insufficient spatio-temporal feature extraction and difficulty in predicting actions in the early execution stage of actions. In our method, GCNs, which have outstanding performance in the field of action recognition, are used to extract the spatio-temporal features of the skeleton. The model learns how to optimize the feature distribution of partial videos from the features of full videos through adversarial learning. Experiments on two challenging action prediction datasets show that our method performs well on skeleton-based early action prediction. State-of-the-art performance is reported in some observation ratios.
