WiFi-based Gesture Recognition on Real-world collected dataset In-Domain
95.18AccuracyCNN-GRU
Evaluation results
| Method | Links | ||
|---|---|---|---|
| CNN-GRUParameters=537,669, GFLOPS=1.892026.03 | 95.18 | 95.18 | |
| ResNet3D-10Parameters=14,358,085, GFLOPS=3.752026.03 | 92.53 | 92.54 | |
| CNN-LSTMParameters=587,077, GFLOPS=1.892026.03 | 91.81 | 91.81 | |
| ResNet3D-18Parameters=33,162,565, GFLOPS=7.152026.03 | 91.81 | 91.86 | |
| ResNet3D-50Parameters=46,165,317, GFLOPS=9.492026.03 | 89.64 | 89.71 | |
| ViViTParameters=144,901, GFLOPS=0.082026.03 | 82.89 | 83.04 | |
| TimeSformerParameters=202,629, GFLOPS=0.082026.03 | 81.45 | 81.36 | |
| 3D CNNParameters=344,581, GFLOPS=3.772026.03 | 78.8 | 78.99 | |
| CNN-RNNParameters=438,853, GFLOPS=1.882026.03 | 77.83 | 78.12 | |
| GesFi2026.01 | 62.89 | — | |
| WiGRUNT2026.01 | 54.44 | — | |
| WiSR2026.01 | 42.67 | — | |
| THAT2026.01 | 41.33 | — | |
| ImgFi2026.01 | 24.67 | — |