SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2203.11283 · CVPR · 2022

NeRFusion: Fusing Radiance Fields for Large-Scale Scene Reconstruction

Hao Su, Xiaoshuai Zhang, Sai Bi, Kalyan Sunkavalli, Zexiang Xu

arXiv · PDF · Open in the Atlas

Code that ran

We have not lifted any functions out of this paper's repositories yet, so there is nothing we have run. The repositories linked to it are listed below.

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Figure 1. We propose a novel method to enable fast reconstruction of volumetric radiance fields of large-scale scenes. Our method uses a novel recurrent network -that is generalizable and trained across scenes -to sequentially reconstruct a radiance field of a large indoor scene in ScanNet [11] from an input image sequence (marked in yellow) via direct inference. The predicted field can be directly used to render realistic images at novel viewpoints (marked in red), achieving comparable quality to NeRF [28] that takes 12 hours per-scene optimization. This radiance field can be further fine-tuned for a short period of 20 min, leading to boosted quality that significantly outperforms NeRF.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2203.11283")
get_code_for_paper("2203.11283")
have("2203.11283")

The run record, dated, one paper per request, free:

curl https://syntology.ai/api/ran/2203.11283.json

A badge for a README (the split and the date, never a ratio):

[![Syntology run record](https://syntology.ai/api/ran/2203.11283.svg)](https://syntology.ai/paper/2203.11283)

Connect an agent — have() is free.