Traffic Sign Recognition on TSRD
97.9AccuracyQDS-SNN
Evaluation results
| Method | Links | ||
|---|---|---|---|
| QDS-SNNArchitecture=QDS-SNN, Time Steps=6, Params(M)=17.77, Train Method=SG, Power(mJ)=2.382026.06 | 97.9 | 97.68 | |
| SpikformerArchitecture=Spikformer-8-768, Time Steps=4, Params(M)=68.73, Train Method=SG, Power(mJ)=23.342026.06 | 97.85 | 97.52 | |
| TETArchitecture=Spiking-ResNet-34, Time Steps=6, Params(M)=23.16, Train Method=SG2026.06 | 97.31 | 96.85 | |
| QDS-SNNArchitecture=QDS-SNN, Time Steps=4, Params(M)=17.77, Train Method=SG, Power(mJ)=1.672026.06 | 97.21 | 96.99 | |
| SA-SCNNArchitecture=SA-SCNN, Time Steps=10, Train Method=SG2026.06 | 97.16 | — | |
| SEW-ResNetArchitecture=SEW-ResNet-34, Time Steps=4, Params(M)=23.16, Train Method=SG, Power(mJ)=4.322026.06 | 97.06 | 96.58 | |
| MS-ResNetArchitecture=MS-ResNet-34, Time Steps=6, Params(M)=23.17, Train Method=SG, Power(mJ)=5.322026.06 | 97.03 | 96.51 | |
| QDS-SNNArchitecture=QDS-SNN, Time Steps=2, Params(M)=17.77, Train Method=SG, Power(mJ)=1.122026.06 | 96.89 | 96.5 | |
| ANN2SNNArchitecture=ResNet-34, Time Steps=64, Params(M)=23.16, Train Method=ANN-SNN2026.06 | 96.12 | — | |
| MS-ResNetArchitecture=MS-ResNet-18, Time Steps=6, Params(M)=12.53, Train Method=SG, Power(mJ)=4.412026.06 | 95.8 | 95.2 | |
| tdBNArchitecture=Spiking-ResNet-34, Time Steps=6, Params(M)=23.16, Train Method=SG, Power(mJ)=6.632026.06 | 95.18 | 94.22 |