Table 2 Reconstruction performance with different methods.
From: Neural encoding with unsupervised spiking convolutional neural network
Dataset | Methods | PCC | SSIM |
|---|---|---|---|
SMLR31 | 0.481 ± 0.096 | 0.191 ± 0.043 | |
DGMM+33 | 0.502 ± 0.193 | 0.360 ± 0.050 | |
Handwritten character | Denoiser GAN34 | 0.319 ± 0.032 | 0.465 ± 0.031 |
CNN | 0.188 ± 0.149 | 0.258 ± 0.073 | |
SCNN (Ours) | 0.561 ± 0.169 | 0.409 ± 0.166 | |
DCCAE-A32 | 0.548 ± 0.044 | 0.358 ± 0.097 | |
DCCAE-S32 | 0.511 ± 0.057 | 0.552 ± 0.088 | |
Handwritten digit | Denoiser GAN34 | 0.531 ± 0.049 | 0.529 ± 0.043 |
CNN | 0.738 ± 0.107 | 0.532 ± 0.122 | |
SCNN (Ours) | 0.762 ± 0.056 | 0.578 ± 0.082 | |
CNN | 0.191 ± 0.271 | 0.344 ± 0.120 | |
Colorful natural image | GNet | 0.166 ± 0.252 | 0.343 ± 0.109 |
SCNN (Ours) | 0.206 ± 0.231 | 0.347 ± 0.116 |
