Attention-based Dual-stream Vision Transformer for Radar Gait Recognition
This work addresses gait recognition for security or healthcare applications by enhancing accuracy in challenging conditions, though it is incremental as it builds on existing representations and Vision Transformer architectures.
The paper tackles radar gait recognition by proposing a dual-stream Vision Transformer with attention-based fusion to combine spectrogram and cadence velocity diagram representations, achieving significant performance improvements over state-of-the-art methods on a large benchmark dataset.
Radar gait recognition is robust to light variations and less infringement on privacy. Previous studies often utilize either spectrograms or cadence velocity diagrams. While the former shows the time-frequency patterns, the latter encodes the repetitive frequency patterns. In this work, a dual-stream neural network with attention-based fusion is proposed to fully aggregate the discriminant information from these two representations. The both streams are designed based on the Vision Transformer, which well captures the gait characteristics embedded in these representations. The proposed method is validated on a large benchmark dataset for radar gait recognition, which shows that it significantly outperforms state-of-the-art solutions.