Improving Neural Indoor Surface Reconstruction with Mask-Guided Adaptive Consistency Constraints
This work addresses a specific bottleneck in neural indoor surface reconstruction for computer vision applications, offering incremental improvements in detail accuracy.
The paper tackles the problem of inaccurate prior estimation in neural implicit surface reconstruction for 3D scenes from 2D images, proposing a two-stage training process with mask-guided adaptive consistency constraints to reduce errors and achieve high-quality reconstructions with rich geometric details.
3D scene reconstruction from 2D images has been a long-standing task. Instead of estimating per-frame depth maps and fusing them in 3D, recent research leverages the neural implicit surface as a unified representation for 3D reconstruction. Equipped with data-driven pre-trained geometric cues, these methods have demonstrated promising performance. However, inaccurate prior estimation, which is usually inevitable, can lead to suboptimal reconstruction quality, particularly in some geometrically complex regions. In this paper, we propose a two-stage training process, decouple view-dependent and view-independent colors, and leverage two novel consistency constraints to enhance detail reconstruction performance without requiring extra priors. Additionally, we introduce an essential mask scheme to adaptively influence the selection of supervision constraints, thereby improving performance in a self-supervised paradigm. Experiments on synthetic and real-world datasets show the capability of reducing the interference from prior estimation errors and achieving high-quality scene reconstruction with rich geometric details.