提出StreamOcc框架，实现高效实时3D占据预测。
Streaming Dense Voxel Representations for 3D Occupancy Prediction
- 先优化传播体素特征，再注入动态物体语义查询
- 在SurroundOcc和Occ3D-nuScenes上刷新性能纪录
- 实时运行仅需83.3毫秒/帧，内存占用2.8GB
本文探索密集体素流用于高精度、高效率的3D占据预测。尽管密集体素表示能提供精细空间细节，流式处理可实现高效时间建模，但直接结合会引入两大挑战：(i) 插值导致的时间对齐扭曲，(ii) 运动错位与图像到体素投影中的细节丢失造成动态物体表征下降。为此，我们提出StreamOcc框架，采用两种聚合策略：首先在时间累积前优化传播体素特征以减少扭曲伪影；随后选择性地将编码动态物体语义的实例级查询特征注入对应占据体素区域，实现时序一致性建模的同时强化动态物体表征。该方法实现有效密集体素流处理，在SurroundOcc基准和Occ3D-nuScenes上达到最先进性能，分别在总体/动态物体mIoU上优于此前最优方法1.3/2.5和1.5/2.0，且每帧仅需83.3毫秒，内存占用2.8 GB。
原文摘要 · Abstract (English)
In this paper, we explore dense voxel streaming for accurate and efficient 3D occupancy prediction. While dense voxel representations offer fine-grained spatial details and streaming paradigm enables efficient temporal processing, naively combining the two introduces key challenges: (i) warping-induced distortions caused by interpolation used for temporal alignment, and (ii) degraded dynamic object representations due to motion misalignment and detail loss in image-to-voxel projection. To address these, we propose StreamOcc, a novel framework that utilizes two aggregation strategies. Specifically, it first refines propagated voxel features to reduce warping artifacts before temporal accumulation, and then selectively injects instance-level query features encoding dynamic-object semantics into the corresponding occupied voxel regions, preserving temporally consistent modeling while strengthening dynamic object representations. Unlocking effective dense voxel streaming, StreamOcc achieves state-of-the-art performance on SurroundOcc-benchmark and Occ3D-nuScenes under real-time constraints, outperforming the prior best methods by +1.3/2.5 and +1.5/2.0 in (overall/dynamic object) mIoU, respectively, while running at 83.3 ms per frame with only 2.8 GB of memory. The project page is available at https://moonseokha.github.io/StreamOcc/.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。