SpaceCache speeds up semantic segmentation of point cloud streams
SpaceCache: Intermediate Feature Reuse for Efficient Semantic Segmentation on Point Cloud Streams
Computer Vision and Pattern Recognition
Summary
Processing 3D point cloud data from sensors is important for self-driving cars and robots, but analyzing every frame fully can be slow. The authors found that parts of the data stay similar across nearby frames. They created SpaceCache, which remembers and reuses features from earlier frames to avoid recalculating everything. This method speeds up processing significantly with only a small drop in accuracy.
What this means in practice
- •For autonomous vehicle engineers: Improve the efficiency of point cloud segmentation pipelines to enable faster perception in real-time driving systems.
- •For robotics system developers: Accelerate 3D environment understanding in robots by reusing intermediate features across sensor frames for quicker response.
Authors
Dongheon Yang, Beomseok Kang, Jae-Joon Kim
Abstract
Point-cloud streams are central to autonomous driving and robotics, yet existing segmentation networks process every frame from scratch, although consecutive frames largely overlap. We observe that the intermediate features of point-cloud transformers remain semantically consistent across frames for the same spatial region, especially in deep stages. Building on this, we present SpaceCache, a training-free feature cache that reuses these features across frames. Tokens of previously observed regions skip a sequence of transformer blocks and take their cached features, only the remaining tokens are computed, and the full token layout is restored for dense prediction. Lightweight cache management keeps reused features reliable by limiting each entry's age and evicting low-confidence entries. Unlike token merging, SpaceCache selects tokens at negligible cost and runs its cache operations once per encoder stage. On both outdoor and indoor benchmarks, such as nuScenes, Waymo, and ScanNet, with two transformer backbones of Utonia and LitePT-B, SpaceCache achieves up to 1.65x speedup with less than 1-point mIoU drop and consistently outperforms token-merging methods.