SparseOcc++: Geometry-Aware Sparse Latent Representation for Semantic Occupancy Prediction
For autonomous driving, this work addresses the computational inefficiency and geometric ambiguity of dense and sparse occupancy prediction methods, offering a faster and more accurate solution.
SparseOcc++ introduces a geometry-aware sparse framework that decouples scene completion from semantic segmentation for 3D occupancy prediction, achieving state-of-the-art results with 2.3 IoU improvement and 3.9x speedup over SparseOcc on nuScenes, and 5.9x speedup over OccFormer on SemanticKITTI.
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.