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.
@article{arxiv.2304.06178,
title = {Dynamic Voxel Grid Optimization for High-Fidelity RGB-D Supervised Surface Reconstruction},
author = {Xiangyu Xu and Lichang Chen and Changjiang Cai and Huangying Zhan and Qingan Yan and Pan Ji and Junsong Yuan and Heng Huang and Yi Xu},
journal= {arXiv preprint arXiv:2304.06178},
year = {2023}
}