DynSUP: Dynamic Gaussian Splatting from An Unposed Image Pair
This addresses the challenge of dynamic scene reconstruction for applications like robotics and AR/VR, where sparse views and unknown poses are common, representing a novel advancement beyond static or multi-view assumptions.
The paper tackles the problem of reconstructing dynamic 3D scenes from only two unposed images, achieving high-fidelity novel view synthesis with accurate temporal consistency and object motion, and significantly outperforms state-of-the-art methods in experiments on synthetic and real-world datasets.
Recent advances in 3D Gaussian Splatting have shown promising results. Existing methods typically assume static scenes and/or multiple images with prior poses. Dynamics, sparse views, and unknown poses significantly increase the problem complexity due to insufficient geometric constraints. To overcome this challenge, we propose a method that can use only two images without prior poses to fit Gaussians in dynamic environments. To achieve this, we introduce two technical contributions. First, we propose an object-level two-view bundle adjustment. This strategy decomposes dynamic scenes into piece-wise rigid components, and jointly estimates the camera pose and motions of dynamic objects. Second, we design an SE(3) field-driven Gaussian training method. It enables fine-grained motion modeling through learnable per-Gaussian transformations. Our method leads to high-fidelity novel view synthesis of dynamic scenes while accurately preserving temporal consistency and object motion. Experiments on both synthetic and real-world datasets demonstrate that our method significantly outperforms state-of-the-art approaches designed for the cases of static environments, multiple images, and/or known poses. Our project page is available at https://colin-de.github.io/DynSUP/.