SparseOcc++: Geometry-Aware Sparse Latent Representation for Semantic Occupancy Prediction

2026-07-06

AI Summary

This work addresses the computational redundancy and geometric ambiguity inherent in existing 3D semantic occupancy prediction methods, which tightly couple scene completion and semantic segmentation. To resolve this, we propose a geometry-aware sparse representation framework that, for the first time, formulates scene completion as signed distance field regression over sparse anchor points. By leveraging orthogonal geometric decomposition and discrete distance learning, our approach explicitly decouples geometric reconstruction from semantic prediction. Furthermore, we introduce a geometry-guided propagation mechanism that restricts semantic reasoning exclusively to geometrically validated regions. Experiments demonstrate that our method achieves a 2.3 percentage point improvement in IoU on nuScenes, while running 3.9× faster than SparseOcc and 5.9× faster than OccFormer on SemanticKITTI.

Abstract

Vision-based 3D semantic occupancy prediction is essential for autonomous driving, yet dense voxel representations waste computation on largely empty space, while BEV and TPV projections compromise fine-grained 3D structure. Fully sparse representations offer an attractive alternative, but existing methods, including SparseOcc, entangle scene completion with semantic prediction by indiscriminately propagating high-dimensional features into empty regions and applying voxel-wise classification. This creates excessive activations, computational overhead, and geometric ambiguity. We present SparseOcc++, a geometry-aware sparse framework that explicitly decouples scene completion from semantic segmentation. SparseOcc++ reformulates completion as signed-distance regression on sparse anchor voxels through a scene completion field (SCF). To model complex outdoor geometry robustly, it combines orthogonal decomposition with discretized distance learning. A geometry-guided propagation module then converts the SCF into a complete volumetric scene and restricts semantic segmentation to geometrically verified regions. Experiments establish new state of the art: SparseOcc++ improves IoU by 2.3 points and is 3.9x faster than SparseOcc on nuScenes, while achieving a 5.9x speedup over OccFormer on SemanticKITTI.

Problem

• semantic occupancy prediction
• sparse representation
• scene completion
• geometry awareness
• autonomous driving

Innovation

• sparse representation
• semantic occupancy prediction
• scene completion
• geometry-aware
• signed-distance regression