SplatSSC: Decoupled Depth-Guided Gaussian Splatting for Semantic Scene Completion

TL;DR

SplatSSC introduces a depth-guided initialization and a Gaussian aggregation method to improve the efficiency and accuracy of monocular 3D semantic scene completion, achieving state-of-the-art results with reduced computational costs.

Contribution

The paper presents a novel depth-guided initialization strategy and a decoupled Gaussian aggregator for monocular SSC, addressing inefficiencies and artifacts of previous object-centric methods.

Findings

Achieves over 6.3% improvement in IoU on Occ-ScanNet

Reduces latency and memory cost by more than 9.3%

Outperforms prior methods with state-of-the-art accuracy

Abstract

Monocular 3D Semantic Scene Completion (SSC) is a challenging yet promising task that aims to infer dense geometric and semantic descriptions of a scene from a single image. While recent object-centric paradigms significantly improve efficiency by leveraging flexible 3D Gaussian primitives, they still rely heavily on a large number of randomly initialized primitives, which inevitably leads to 1) inefficient primitive initialization and 2) outlier primitives that introduce erroneous artifacts. In this paper, we propose SplatSSC, a novel framework that resolves these limitations with a depth-guided initialization strategy and a principled Gaussian aggregator. Instead of random initialization, SplatSSC utilizes a dedicated depth branch composed of a Group-wise Multi-scale Fusion (GMF) module, which integrates multi-scale image and depth features to generate a sparse yet representative set of initial Gaussian primitives. To mitigate noise from outlier primitives, we develop the Decoupled Gaussian Aggregator (DGA), which enhances robustness by decomposing geometric and semantic predictions during the Gaussian-to-voxel splatting process. Complemented with a specialized Probability Scale Loss, our method achieves state-of-the-art performance on the Occ-ScanNet dataset, outperforming prior approaches by over 6.3% in IoU and 4.1% in mIoU, while reducing both latency and memory cost by more than 9.3%.