AI & ML New Capability

Enables online, incremental 3D Gaussian Splatting for thousands of frames by replacing global reprocessing with a causal, streaming update framework.

arXiv · March 17, 2026 · 2603.14232

Renhe Zhang, Yuyang Tan, Jingyu Gong, Zhizhong Zhang, Lizhuang Ma, Yuan Xie, Xin Tan

The Takeaway

Most 3DGS methods are offline and fail at around 80 frames due to rapid memory growth. S2GS scales to 1,000+ frames with significantly slower memory growth, enabling real-time semantic mapping and reconstruction for long-horizon robotics and AR applications.

From the abstract

Existing offline feed-forward methods for joint scene understanding and reconstruction on long image streams often repeatedly perform global computation over an ever-growing set of past observations, causing runtime and GPU memory to increase rapidly with sequence length and limiting scalability. We propose Streaming Semantic Gaussian Splatting (S2GS), a strictly causal, incremental 3D Gaussian semantic field framework: it does not leverage future frames and continuously updates scene geometry,