Motion Feature Augmented Recurrent Neural Network for Skeleton-based Dynamic Hand Gesture Recognition
This addresses gesture recognition for human-computer interaction, but it appears incremental as it builds on existing skeleton-based and RNN approaches with added motion features.
The paper tackled dynamic hand gesture recognition by proposing a motion feature augmented recurrent neural network that extracts finger and global motion features from skeleton sequences, feeding them into a bidirectional RNN to improve classification, and it outperformed state-of-the-art methods.
Dynamic hand gesture recognition has attracted increasing interests because of its importance for human computer interaction. In this paper, we propose a new motion feature augmented recurrent neural network for skeleton-based dynamic hand gesture recognition. Finger motion features are extracted to describe finger movements and global motion features are utilized to represent the global movement of hand skeleton. These motion features are then fed into a bidirectional recurrent neural network (RNN) along with the skeleton sequence, which can augment the motion features for RNN and improve the classification performance. Experiments demonstrate that our proposed method is effective and outperforms start-of-the-art methods.