Depth Reconstruction with Neural Signed Distance Fields in Structured Light Systems
This work addresses depth reconstruction for structured light systems, offering an incremental improvement by isolating geometry optimization with known radiance fields.
The paper tackles depth estimation in multi-frame structured light systems by using neural signed distance fields (SDFs) trained with self-supervised differentiable rendering, achieving superior geometric performance in few-shot scenarios and comparable results with more patterns.
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.