Semantic Segmentation Accuracy Metrics on Cityscapes
0.85Pixel Accpix2pix
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| pix2pixLoss=pix2pix, Supervision=Paired/Supervised2017.03 | 0.85 | 0.4 | 0.32 | — | |
| MSPC2022.03 | 0.74 | 0.296 | — | 0.226 | |
| CUTGAN2022.03 | 0.681 | 0.243 | — | 0.172 | |
| MSPC2022.03 | 0.612 | 0.214 | — | 0.156 | |
| CycleGAN2022.03 | 0.595 | 0.234 | — | 0.171 | |
| CUTGAN2022.03 | 0.587 | 0.225 | — | 0.166 | |
| GCGAN2022.03 | 0.583 | 0.201 | — | 0.128 | |
| CycleGANLoss=CycleGAN (ours)2017.03 | 0.58 | 0.22 | 0.16 | — | |
| GCGAN2022.03 | 0.563 | 0.195 | — | 0.143 | |
| CycleGAN2022.03 | 0.508 | 0.184 | — | 0.117 | |
| Feature loss + GANLoss=Feature loss + GAN2017.03 | 0.5 | 0.1 | 0.06 | — | |
| VAT Modified2022.03 | 0.484 | 0.1 | — | 0.064 | |
| SimGANLoss=SimGAN2017.03 | 0.47 | 0.11 | 0.07 | — | |
| MT Modified2022.03 | 0.455 | 0.145 | — | 0.086 | |
| COGANLoss=COGAN2017.03 | 0.45 | 0.11 | 0.08 | — | |
| BIGAN/ALILoss=BIGAN/ALI2017.03 | 0.41 | 0.13 | 0.07 | — | |
| VAT Modified2022.03 | 0.281 | 0.109 | — | 0.053 | |
| MT Modified2022.03 | 0.121 | 0.055 | — | 0.018 |