Spatial-TTT: Streaming Visual-based Spatial Intelligence with Test-Time Training

TL;DR

Spatial-TTT introduces a novel streaming visual spatial intelligence model that adapts parameters at test time, effectively capturing and organizing long-term spatial information from video streams for improved spatial understanding.

Contribution

The paper proposes a hybrid architecture with test-time training and a spatial-predictive mechanism, along with a new dataset, to enhance long-horizon spatial reasoning in videos.

Findings

Achieves state-of-the-art results on video spatial benchmarks.

Effectively captures geometric and temporal continuity across frames.

Improves long-term spatial understanding in streaming video data.

Abstract

Humans perceive and understand real-world spaces through a stream of visual observations. Therefore, the ability to streamingly maintain and update spatial evidence from potentially unbounded video streams is essential for spatial intelligence. The core challenge is not simply longer context windows but how spatial information is selected, organized, and retained over time. In this paper, we propose Spatial-TTT towards streaming visual-based spatial intelligence with test-time training (TTT), which adapts a subset of parameters (fast weights) to capture and organize spatial evidence over long-horizon scene videos. Specifically, we design a hybrid architecture and adopt large-chunk updates parallel with sliding-window attention for efficient spatial video processing. To further promote spatial awareness, we introduce a spatial-predictive mechanism applied to TTT layers with 3D spatiotemporal convolution, which encourages the model to capture geometric correspondence and temporal continuity across frames. Beyond architecture design, we construct a dataset with dense 3D spatial descriptions, which guides the model to update its fast weights to memorize and organize global 3D spatial signals in a structured manner. Extensive experiments demonstrate that Spatial-TTT improves long-horizon spatial understanding and achieves state-of-the-art performance on video spatial benchmarks. Project page: https://liuff19.github.io/Spatial-TTT.