Novel View Synthesis on 360 dataset multiscale 4x scale factor benchmark leaderboard · SOTA2 Research

Novel View Synthesis on 360 dataset multiscale 4x scale factor

31.57PSNRZip-NeRF

Evaluation Results

Method
Zip-NeRF2023.04	31.57	0.933	0.056
Zip-NeRF (Unscented Multisampling)2023.04	31.57	0.933	0.056
Zip-NeRF (No Appended Scale w)2023.04	31.48	0.931	0.057
Zip-NeRF (No Multisampling)2023.04	31.33	0.927	0.061
Zip-NeRF (No New Interlevel Loss)2023.04	31.31	0.932	0.056
Zip-NeRF (No Downweighting)2023.04	31.25	0.928	0.06
Zip-NeRF (Random Multisampling)2023.04	31.18	0.928	0.061
Zip-NeRF (Un-Normalized Weight Decay)2023.04	31.1	0.93	0.058
mip-NeRF 3602023.04	30.45	0.912	0.077
Zip-NeRF (Small View-Dependent MLP)2023.04	30.32	0.924	0.065
Zip-NeRF (No Weight Decay)2023.04	30.29	0.921	0.067
Zip-NeRF (Jittered)2023.04	29.45	0.893	0.09
Zip-NeRF (Naive Sampling)2023.04	29.24	0.887	0.094
Zip-NeRF (Jittered Supersampling (6x))2023.04	28.91	0.896	0.086
Zip-NeRF (Naive Supersampling (6x))2023.04	28.42	0.881	0.097
mip-NeRF 360 + iNGP2023.04	27.67	0.866	0.116
Instant NGP2023.04	26.84	0.809	0.142