DINO_4D: Semantic-Aware 4D Reconstruction
For researchers in computer vision and robotics, DINO_4D provides a method to integrate semantic awareness into 4D reconstruction, improving performance on dynamic scenes.
DINO_4D introduces frozen DINOv3 features as structural priors for 4D reconstruction of dynamic scenes, suppressing semantic drift during dynamic tracking. It achieves improved tracking accuracy and reconstruction completeness on Point Odyssey and TUM-Dynamics benchmarks while maintaining linear time complexity.
In the intersection of computer vision and robotic perception, 4D reconstruction of dynamic scenes serve as the critical bridge connecting low-level geometric sensing with high-level semantic understanding. We present DINO\_4D, introducing frozen DINOv3 features as structural priors, injecting semantic awareness into the reconstruction process to effectively suppress semantic drift during dynamic tracking. Experiments on the Point Odyssey and TUM-Dynamics benchmarks demonstrate that our method maintains the linear time complexity $O(T)$ of its predecessors while significantly improving Tracking Accuracy (APD) and Reconstruction Completeness. DINO\_4D establishes a new paradigm for constructing 4D World Models that possess both geometric precision and semantic understanding.