Cross-domain gesture recognition method based on diffusion model

In recent years, radar sensor-based gesture recognition has attracted considerable attention; however, it faces challenges such as limited training data and difficulties in generalizing to new environments. To address these challenges, this study proposes a two-stage domain adaptation approach that integrates diffusion model-based data augmentation with a sophisticated domain adaptation network for gesture recognition. To validate the effectiveness of our method, we conducted radar measurement experiments. The results demonstrate that the proposed method significantly improves recognition accuracy, underscoring its great capability for radar-based gesture recognition.