Dynamic Voxel Grid Optimization for High-Fidelity RGB-D Supervised Surface Reconstruction
This work addresses the challenge of efficient and detailed 3D reconstruction for computer vision applications, representing an incremental improvement over existing voxel grid methods.
The paper tackles the problem of high-fidelity 3D surface reconstruction from RGB-D data by introducing a dynamic voxel grid optimization method that adaptively allocates finer voxels to complex regions, resulting in high-quality reconstructions with fine details and substantially faster performance than the baseline NeuralRGBD.
Direct optimization of interpolated features on multi-resolution voxel grids has emerged as a more efficient alternative to MLP-like modules. However, this approach is constrained by higher memory expenses and limited representation capabilities. In this paper, we introduce a novel dynamic grid optimization method for high-fidelity 3D surface reconstruction that incorporates both RGB and depth observations. Rather than treating each voxel equally, we optimize the process by dynamically modifying the grid and assigning more finer-scale voxels to regions with higher complexity, allowing us to capture more intricate details. Furthermore, we develop a scheme to quantify the dynamic subdivision of voxel grid during optimization without requiring any priors. The proposed approach is able to generate high-quality 3D reconstructions with fine details on both synthetic and real-world data, while maintaining computational efficiency, which is substantially faster than the baseline method NeuralRGBD.