Gesture Recognition on DVSGesture
0.993Top-1 AccuracySpike-driven Transformer
Evaluation results
| Method | Links | |
|---|---|---|
| Spike-driven TransformerType=Direct Training, Architecture=Spike-driven Transformer-2-256, #Param (M)=2.55, T=162024.03 | 0.993 | |
| SpikingResformerType=Direct Training, Architecture=SpikingResformer-XTi*, #Param (M)=2.71, T=162024.03 | 0.986 | |
| CF-SNN (Ours)Architecture=VGG, Time Steps=102026.05 | 0.9833 | |
| SpikingformerType=Direct Training, Architecture=Spikingformer-2-256, #Param (M)=2.55, T=162024.03 | 0.983 | |
| TCJAArchitecture=TCJA-SNN, Time Steps=102026.05 | 0.982 | |
| SpikformerType=Direct Training, Architecture=Spikformer-2-256, #Param (M)=2.55, T=162024.03 | 0.979 | |
| PLIFType=Direct Training, Architecture=6 Conv, 2 FC, #Param (M)=1.69, T=202024.03 | 0.9757 | |
| PLIFArchitecture=PLIF Net, Time Steps=202026.05 | 0.9757 | |
| STBP-tdBNType=Direct Training, Architecture=ResNet-17, #Param (M)=1.40, T=402024.03 | 0.9687 | |
| EventMixArchitecture=ResNet-18, Time Steps=102026.05 | 0.9675 | |
| SpikingformerType=Direct Training, Architecture=Spikingformer-2-256, #Param (M)=2.55, T=102024.03 | 0.962 | |
| SpikformerType=Direct Training, Architecture=Spikformer-2-256, #Param (M)=2.55, T=102024.03 | 0.958 | |
| SpikingResformerType=Transfer Learning, Architecture=SpikingResformer-Ti, #Param (M)=10.76, T=102024.03 | 0.934 | |
| SpikingResformerType=Transfer Learning, Architecture=SpikingResformer-S, #Param (M)=17.25, T=102024.03 | 0.934 | |
| nD-sincos2026.02 | 0.934 | |
| PaPE2026.02 | 0.934 | |
| RoPE2026.02 | 0.931 | |
| PaPE-RIrotation-invariant=true2026.02 | 0.906 | |
| nD-ALiBi2026.02 | 0.892 | |
| RoPE-Mixed2026.02 | 0.833 |