3D shape reconstruction on UHM12k (test)
0.0797Mean Euclidean DistanceLAMM-MLPMixer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| LAMM-MLPMixerlatent size=256, training regime=autoencoding, backbone=MLPMixer2024.01 | 0.0797 | — | |
| LAMM-Transformerlatent size=256, training regime=autoencoding, backbone=Transformer2024.01 | 0.0908 | — | |
| PCAlatent size=256, training regime=autoencoding2024.01 | 0.1042 | — | |
| SplineNet++latent size=256, training regime=autoencoding2024.01 | 0.1153 | — | |
| Neural3DMMlatent size=256, training regime=autoencoding2024.01 | 0.1182 | — | |
| COMAlatent size=256, training regime=autoencoding2024.01 | 0.1311 | — | |
| COMAlatent size=256, training regime=autoencoding2024.01 | — | 15.4 | |
| LAMM-MLPMixerlatent size=256, training regime=autoencoding, backbone=MLPMixer2024.01 | — | 11.48 | |
| LAMM-Transformerlatent size=256, training regime=autoencoding, backbone=Transformer2024.01 | — | 13.7 | |
| Neural3DMMlatent size=256, training regime=autoencoding2024.01 | — | 13.88 | |
| PCAlatent size=256, training regime=autoencoding2024.01 | — | 11.73 | |
| SplineNet++latent size=256, training regime=autoencoding2024.01 | — | 13.55 |