SAM4D is a multi-modal and temporal foundation model for segmentation in autonomous driving using Unified Multi-modal Positional Encoding and Motion-aware Cross-modal Memory Attention, with a multi-modal automated data engine generating pseudo-labels.
We present SAM4D, a multi-modal and temporal foundation model designed for
promptable segmentation across camera and LiDAR streams. Unified Multi-modal
Positional Encoding (UMPE) is introduced to align camera and LiDAR features in
a shared 3D space, enabling seamless cross-modal prompting and interaction.
Additionally, we propose Motion-aware Cross-modal Memory Attention (MCMA),
which leverages ego-motion compensation to enhance temporal consistency and
long-horizon feature retrieval, ensuring robust segmentation across dynamically
changing autonomous driving scenes. To avoid annotation bottlenecks, we develop
a multi-modal automated data engine that synergizes VFM-driven video masklets,
spatiotemporal 4D reconstruction, and cross-modal masklet fusion. This
framework generates camera-LiDAR aligned pseudo-labels at a speed orders of
magnitude faster than human annotation while preserving VFM-derived semantic
fidelity in point cloud representations. We conduct extensive experiments on
the constructed Waymo-4DSeg, which demonstrate the powerful cross-modal
segmentation ability and great potential in data annotation of proposed SAM4D.