MIPS-Fusion: Multi-Implicit-Submaps for Scalable and Robust Online Neural RGB-D Reconstruction

TL;DR

MIPS-Fusion introduces a scalable, robust neural RGB-D reconstruction method using multi-implicit-submaps, enabling efficient large-scale scene modeling and fast camera motion handling through novel divide-and-conquer neural representations and hybrid tracking.

Contribution

It presents a novel multi-implicit-submap neural representation and hybrid tracking approach, improving scalability, robustness, and efficiency over existing neural RGB-D reconstruction methods.

Findings

Achieves higher reconstruction quality than state-of-the-art methods.

Handles fast camera motions effectively.

Supports large-scale scene reconstruction.

Abstract

We introduce MIPS-Fusion, a robust and scalable online RGB-D reconstruction method based on a novel neural implicit representation -- multi-implicit-submap. Different from existing neural RGB-D reconstruction methods lacking either flexibility with a single neural map or scalability due to extra storage of feature grids, we propose a pure neural representation tackling both difficulties with a divide-and-conquer design. In our method, neural submaps are incrementally allocated alongside the scanning trajectory and efficiently learned with local neural bundle adjustments. The submaps can be refined individually in a back-end optimization and optimized jointly to realize submap-level loop closure. Meanwhile, we propose a hybrid tracking approach combining randomized and gradient-based pose optimizations. For the first time, randomized optimization is made possible in neural tracking with several key designs to the learning process, enabling efficient and robust tracking even under fast camera motions. The extensive evaluation demonstrates that our method attains higher reconstruction quality than the state of the arts for large-scale scenes and under fast camera motions.