Matching-Free Depth Recovery from Structured Light

TL;DR

This paper presents a new matching-free depth recovery method from structured light images using a density voxel grid and self-supervised volume rendering, achieving faster convergence and improved accuracy over existing techniques.

Contribution

It introduces a novel matching-free depth estimation approach employing a density voxel grid and self-supervised differentiable volume rendering, enhancing speed and geometric accuracy.

Findings

Achieves ~30% reduction in depth errors compared to matching-based methods.

Approximately three times faster training than previous implicit matching-free methods.

Outperforms existing techniques in synthetic and real-world scenes.

Abstract

We introduce a novel approach for depth estimation using images obtained from monocular structured light systems. In contrast to many existing methods that depend on image matching, our technique employs a density voxel grid to represent scene geometry. This grid is trained through self-supervised differentiable volume rendering. Our method leverages color fields derived from the projected patterns in structured light systems during the rendering process, facilitating the isolated optimization of the geometry field. This innovative approach leads to faster convergence and high-quality results. Additionally, we integrate normalized device coordinates (NDC), a distortion loss, and a distinctive surface-based color loss to enhance geometric fidelity. Experimental results demonstrate that our method outperforms current matching-based techniques in terms of geometric performance in few-shot scenarios, achieving an approximately 30% reduction in average estimated depth errors for both synthetic scenes and real-world captured scenes. Moreover, our approach allows for rapid training, being approximately three times faster than previous matching-free methods that utilize implicit representations.