Graph Convolutional Networks (GCNs) and Transformers have been widely applied to skeleton-based human action recognition, with each offering unique advantages in capturing spatial relationships and long-range dependencies. However, for most GCN methods, the construction of topological structures relies solely on the spatial information of human joints, limiting their ability to directly capture richer spatio-temporal dependencies. Additionally, the self-attention modules of many Transformer methods lack topological structure information, restricting the robustness and generalization of the models. To address these issues, we propose a Joint Trajectory Graph (JTG) that integrates spatio-temporal information into a uniform graph structure. We also present a Joint Trajectory GraphFormer (JT-GraphFormer), which directly captures the spatio-temporal relationships among all joint trajectories for human action recognition. To better integrate topological information into spatio-temporal relationships, we introduce a Spatio-Temporal Dijkstra Attention (STDA) mechanism to calculate relationship scores for all the joints in JTG. Furthermore, we incorporate the Koopman operator into the classification stage to enhance the model's representation ability and classification performance. Experiments demonstrate that JT-GraphFormer achieves outstanding performance in human action recognition tasks, outperforming state-of-the-art methods on the NTU RGB+D, NTU RGB+D 120, and N-UCLA datasets.
1 November 2020 in Journal of Physics: Conference Series
Convolutional Neural Network (CNN) has been applied to various tasks with great success. Adding an activation function is an important way to introduce nonlinearity into convolutional neural networks. The commonly used activation functions mostly use different forms of negative feedback for the negative input, however, some researchers have recently proposed positive feedback methods for the negative input such as Concatenated Rectified Linear Units (CReLU) and Linearly Scaled Hyperbolic Tangent (LiSHT) and achieved better performance. To explore this idea further more, we propose a new activation function called Difference Exponentially Linear Unit (DELU). The proposed DELU activation function can optionally provide positive and negative feedback for different values of negative inputs. Our experimental results on the commonly used datasets such as Fashion Mnist, CIFAR10, and Imagenet show that DELU outperforms other six activation functions, Leaky ReLU, ReLU, ELU, SELU, Swish, SERLU.