Depth Reconstruction with Neural Signed Distance Fields in Structured Light Systems

TL;DR

This paper presents a neural implicit depth estimation method using signed distance fields tailored for structured light systems, achieving high geometric accuracy with limited data and fixed device setup.

Contribution

It introduces a neural SDF-based depth reconstruction approach that leverages known radiance in structured light, enabling efficient and accurate geometry estimation.

Findings

Outperforms existing methods in geometric accuracy with few-shot data

Achieves comparable results with more pattern data

Ensures convergence with fixed device positioning

Abstract

We introduce a novel depth estimation technique for multi-frame structured light setups using neural implicit representations of 3D space. Our approach employs a neural signed distance field (SDF), trained through self-supervised differentiable rendering. Unlike passive vision, where joint estimation of radiance and geometry fields is necessary, we capitalize on known radiance fields from projected patterns in structured light systems. This enables isolated optimization of the geometry field, ensuring convergence and network efficacy with fixed device positioning. To enhance geometric fidelity, we incorporate an additional color loss based on object surfaces during training. Real-world experiments demonstrate our method's superiority in geometric performance for few-shot scenarios, while achieving comparable results with increased pattern availability.