Neural radiance fields (NeRFs) enable novel view synthesis with unprecedented visual quality. However, to render photorealistic images, NeRFs require hundreds of deep multilayer perceptron (MLP) evaluations - for each pixel. This is prohibitively expensive and makes real-time rendering infeasible, even on powerful modern GPUs. In this paper, we propose a novel approach to distill and bake NeRFs into highly efficient mesh-based neural representations that are fully compatible with the massively parallel graphics rendering pipeline. We represent scenes as neural radiance features encoded on a two-layer duplex mesh, which effectively overcomes the inherent inaccuracies in 3D surface reconstruction by learning the aggregated radiance information from a reliable interval of ray-surface intersections. To exploit local geometric relationships of nearby pixels, we leverage screen-space convolutions instead of the MLPs used in NeRFs to achieve high-quality appearance. Finally, the performance of the whole framework is further boosted by a novel multi-view distillation optimization strategy. We demonstrate the effectiveness and superiority of our approach via extensive experiments on a range of standard datasets.
@article{arxiv.2304.10537,
title = {Learning Neural Duplex Radiance Fields for Real-Time View Synthesis},
author = {Ziyu Wan and Christian Richardt and Aljaž Božič and Chao Li and Vijay Rengarajan and Seonghyeon Nam and Xiaoyu Xiang and Tuotuo Li and Bo Zhu and Rakesh Ranjan and Jing Liao},
journal= {arXiv preprint arXiv:2304.10537},
year = {2023}
}