Mip-NeRF 360
Mip-NeRF 360 is a collection of four indoor and five outdoor object-centric scenes. The camera trajectory is an orbit around the object with fixed elevation and radius. The test set takes each n-th frame of the trajectory as test views.
Authors: Jonathan T. Barron, Ben Mildenhall, Dor Verbin, Pratul P. Srinivasan, Peter Hedman
Method | PSNR | SSIM | LPIPS | Time | GPU Mem. | |
---|---|---|---|---|---|---|
Zip-NeRF | 5h 30m 49s | 26.19 GB | ||||
Mip-NeRF 360 | 7h 29m 42s | 126.99 GB | ||||
Mip-Splatting | 25m 1s | 10.96 GB | ||||
Gaussian Splatting | 22m 45s | 11.12 GB | ||||
NerfStudio | 19m 50s | 3.80 GB | ||||
Tetra-NeRF | 17h 32m 35s | 13.37 GB | ||||
Instant NGP | 4m 16s | 5.62 GB |
PSNR
Peak Signal to Noise Ratio. The higher the better.
Method | garden | bicycle | flowers | treehill | stump | kitchen | bonsai | counter | room |
---|---|---|---|---|---|---|---|---|---|
Zip-NeRF | |||||||||
Mip-NeRF 360 | |||||||||
Mip-Splatting | |||||||||
Gaussian Splatting | |||||||||
NerfStudio | |||||||||
Tetra-NeRF | |||||||||
Instant NGP |
SSIM
Structural Similarity Index. The higher the better. The implementation matches JAX's SSIM and torchmetrics's SSIM (with default parameters).
Method | garden | bicycle | flowers | treehill | stump | kitchen | bonsai | counter | room |
---|---|---|---|---|---|---|---|---|---|
Zip-NeRF | |||||||||
Mip-Splatting | |||||||||
Gaussian Splatting | |||||||||
Mip-NeRF 360 | |||||||||
NerfStudio | |||||||||
Instant NGP | |||||||||
Tetra-NeRF |
LPIPS
Learned Perceptual Image Patch Similarity. The lower the better. The implementation uses VGG backbone and matches lpips pip package with checkpoint version 0.1
Method | garden | bicycle | flowers | treehill | stump | kitchen | bonsai | counter | room |
---|---|---|---|---|---|---|---|---|---|
Zip-NeRF | |||||||||
Gaussian Splatting | |||||||||
Mip-Splatting | |||||||||
Mip-NeRF 360 | |||||||||
NerfStudio | |||||||||
Tetra-NeRF | |||||||||
Instant NGP |