Any4D: Unified Feed-Forward Metric 4D Reconstruction
This work addresses the challenge of accurate and efficient 4D scene reconstruction for applications in robotics and computer vision, representing a strong specific gain rather than a foundational advancement.
The paper tackles the problem of metric-scale, dense 4D reconstruction from multi-view inputs by introducing Any4D, a scalable transformer that directly predicts per-pixel motion and geometry for N frames, achieving 2-3 times lower error and 15 times faster computation compared to prior methods.
We present Any4D, a scalable multi-view transformer for metric-scale, dense feed-forward 4D reconstruction. Any4D directly generates per-pixel motion and geometry predictions for N frames, in contrast to prior work that typically focuses on either 2-view dense scene flow or sparse 3D point tracking. Moreover, unlike other recent methods for 4D reconstruction from monocular RGB videos, Any4D can process additional modalities and sensors such as RGB-D frames, IMU-based egomotion, and Radar Doppler measurements, when available. One of the key innovations that allows for such a flexible framework is a modular representation of a 4D scene; specifically, per-view 4D predictions are encoded using a variety of egocentric factors (depthmaps and camera intrinsics) represented in local camera coordinates, and allocentric factors (camera extrinsics and scene flow) represented in global world coordinates. We achieve superior performance across diverse setups - both in terms of accuracy (2-3X lower error) and compute efficiency (15X faster), opening avenues for multiple downstream applications.