Expansive Supervision for Neural Radiance Field

TL;DR

This paper introduces Expansive Supervision, a method that reduces training time and memory usage in Neural Radiance Fields by selectively supervising crucial pixels, enabling faster and more efficient rendering without quality loss.

Contribution

The paper proposes a novel supervision technique based on partial ray selection that significantly decreases computational costs in NeRF training while preserving visual quality.

Findings

52% memory savings during training

16% reduction in training time

Maintains comparable visual quality to standard methods

Abstract

Neural Radiance Field (NeRF) has achieved remarkable success in creating immersive media representations through its exceptional reconstruction capabilities. However, the computational demands of dense forward passes and volume rendering during training continue to challenge its real-world applications. In this paper, we introduce Expansive Supervision to reduce time and memory costs during NeRF training from the perspective of partial ray selection for supervision. Specifically, we observe that training errors exhibit a long-tail distribution correlated with image content. Based on this observation, our method selectively renders a small but crucial subset of pixels and expands their values to estimate errors across the entire area for each iteration. Compared to conventional supervision, our approach effectively bypasses redundant rendering processes, resulting in substantial reductions in both time and memory consumption. Experimental results demonstrate that integrating Expansive Supervision within existing state-of-the-art acceleration frameworks achieves 52% memory savings and 16% time savings while maintaining comparable visual quality.