SparseOcc++: Geometry-Aware Sparse Latent Representation for Semantic Occupancy Prediction
2026-07-06 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors improved a method for predicting 3D maps from camera images used in self-driving cars. They noticed that previous approaches mixed guessing the shape of the scene and labeling what things are, which slowed down the process and caused confusion. Their method, SparseOcc++, separates these tasks by first estimating the shape using a special technique, then labeling only the confirmed parts. This makes their system faster and more accurate compared to earlier methods on popular driving datasets.
3D semantic occupancy predictionsparse voxel representationscene completionsemantic segmentationsigned-distance regressionvolumetric sceneorthogonal decompositionnuScenes datasetSemanticKITTI dataset
Authors
Pin Tang, Zhongdao Wang, Guoqing Wang, Xiangxuan Ren, Chao Ma
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.