Comparisons with Baselines on Mip-NeRF360

Here we show rendered views from our full model, trained on DL3DV and evaluated zero-shot on Mip-NeRF360, against 3DGS which is optimized per scene for ~13 minutes.

Hover over the labels at the left to select which output to view. Hold mouse down anywhere to flip to ground truth for comparison.

Ground truth
LVTSH-rgba
3DGS
LVTSH-rgba depth