Novel View Synthesis on ZJU-MoCap (source models' seen poses)
31.57PSNRNeural Human Performer
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Neural Human PerformerProtocol=per-scene optimization2021.09 | 31.57 | 0.966 | |
| Neural Human PerformerTraining source=source models, Optimization Strategy=per-scene2021.09 | 31.57 | 0.966 | |
| Neural Human PerformerTraining=generalizable, Protocol=trained on source models2021.09 | 28.73 | 0.936 | |
| Neural Human PerformerTraining source=source models, Optimization Strategy=cross-subject2021.09 | 28.73 | 0.936 | |
| Neural BodyTraining=per-person, Protocol=trained on source models2021.09 | 28.51 | 0.947 | |
| Neural BodyTraining source=source models, Optimization Strategy=per-person2021.09 | 28.51 | 0.947 | |
| Neural Human RenderingTraining=per-person, Protocol=trained on source models2021.09 | 23.95 | 0.897 | |
| Neural Human RenderingTraining source=source models, Optimization Strategy=per-person2021.09 | 23.95 | 0.897 | |
| Neural TexturesTraining=per-person, Protocol=trained on source models2021.09 | 23.86 | 0.896 | |
| Neural TexturesTraining source=source models, Optimization Strategy=per-person2021.09 | 23.86 | 0.896 |