CASA-SDF: Curriculum-Aware Spatial Adaptation with Curvature-Guided Density for Neural Implicit Surface Reconstruction

2026-07-15

AI Summary

This work addresses the challenge of high-fidelity indoor scene reconstruction posed by geometric heterogeneity: large textureless planar regions require strong regularization to suppress artifacts, while thin, intricate structures are prone to oversmoothing due to the spectral bias of MLPs. To reconcile this trade-off, the authors propose CASA-SDF, a unified framework that balances smoothness and detail preservation through a cooperative spatially adaptive strategy. The method innovatively integrates pixel-wise curriculum learning with curvature-guided local density transformation—employing mixed spatially adaptive uncertainty annealing (SAUA) that leverages semantic and photometric uncertainties to construct a supervision curriculum, and curvature-aware locally adaptive density transformation (CALADT) that dynamically adjusts the sharpness of the SDF-to-density mapping. Experiments demonstrate that CASA-SDF significantly improves surface completeness and high-frequency detail recovery across multiple indoor benchmarks while maintaining stability in large planar regions.

Abstract

Neural implicit representations have emerged as a powerful paradigm for 3D reconstruction. However, high-fidelity indoor surface reconstruction remains a significant challenge, primarily due to the pronounced \emph{geometric heterogeneity} of indoor scenes. Large texture-less planar regions typically require stronger regularization to suppress high-frequency artifacts, while thin structures demand sharper, more adaptive representations to mitigate the spectral bias of multi-layer perceptrons (MLPs) and prevent over-smoothing. Existing approaches often rely on spatially indiscriminate prior supervision and a scene-global SDF-to-density transformation, which constrains their ability to balance planar smoothness and detail preservation. In this paper, we propose CASA-SDF (Curriculum-Aware Spatial Adaptation for SDF), a unified framework that addresses this challenge via complementary adaptations of supervision and representation capacity. Specifically, Hybrid Spatially-Adaptive Uncertainty Annealing (SAUA) fuses semantic and photometric uncertainties to construct a pixel-wise curriculum for monocular prior supervision. This strategy maintains regularization in reliable regions while attenuating unreliable supervision early in training to enable data-driven photometric refinement. Meanwhile, Curvature-Aware Locally Adaptive Density Transformation (CALADT) progressively modulates the sharpness of the SDF-to-density mapping via a curvature proxy to enhance the representation of thin structures. Extensive experiments on benchmark indoor datasets demonstrate that CASA-SDF improves surface completeness and detail recovery on high-frequency structures, without compromising the stability of planar surfaces.

Problem

• geometric heterogeneity
• neural implicit surface reconstruction
• spectral bias
• SDF-to-density transformation
• high-fidelity indoor reconstruction

Innovation

• Spatially-Adaptive Supervision
• Curvature-Guided Density
• Neural Implicit Surface
• Curriculum Learning
• SDF-to-Density Transformation