Easi3R: Estimating Disentangled Motion from DUSt3R Without Training

• URL: http://arxiv.org/abs/2503.24391v1
• Date: Mon, 31 Mar 2025 17:59:58 GMT
• Title: Easi3R: Estimating Disentangled Motion from DUSt3R Without Training
• Authors: Xingyu Chen, Yue Chen, Yuliang Xiu, Andreas Geiger, Anpei Chen,
• Abstract summary: We introduce Easi3R, a simple yet efficient training-free method for 4D reconstruction.<n>Our approach applies attention adaptation during inference, eliminating the need for from-scratch pre-training or network fine-tuning.<n>Our experiments on real-world dynamic videos demonstrate that our lightweight attention adaptation significantly outperforms previous state-of-the-art methods.
• Score: 48.87063562819018
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Recent advances in DUSt3R have enabled robust estimation of dense point clouds and camera parameters of static scenes, leveraging Transformer network architectures and direct supervision on large-scale 3D datasets. In contrast, the limited scale and diversity of available 4D datasets present a major bottleneck for training a highly generalizable 4D model. This constraint has driven conventional 4D methods to fine-tune 3D models on scalable dynamic video data with additional geometric priors such as optical flow and depths. In this work, we take an opposite path and introduce Easi3R, a simple yet efficient training-free method for 4D reconstruction. Our approach applies attention adaptation during inference, eliminating the need for from-scratch pre-training or network fine-tuning. We find that the attention layers in DUSt3R inherently encode rich information about camera and object motion. By carefully disentangling these attention maps, we achieve accurate dynamic region segmentation, camera pose estimation, and 4D dense point map reconstruction. Extensive experiments on real-world dynamic videos demonstrate that our lightweight attention adaptation significantly outperforms previous state-of-the-art methods that are trained or finetuned on extensive dynamic datasets. Our code is publicly available for research purpose at https://easi3r.github.io/

Related papers

• Geo4D: Leveraging Video Generators for Geometric 4D Scene Reconstruction [72.54905331756076]
  We introduce Geo4D, a method to repurpose video diffusion models for monocular 3D reconstruction of dynamic scenes. By leveraging the strong dynamic prior captured by such video models, Geo4D can be trained using only synthetic data.
  arXiv  Detail & Related papers  (2025-04-10T17:59:55Z)
• Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos [76.07894127235058]
  We present a system for mining high-quality 4D reconstructions from internet stereoscopic, wide-angle videos.<n>We use this method to generate large-scale data in the form of world-consistent, pseudo-metric 3D point clouds.<n>We demonstrate the utility of this data by training a variant of DUSt3R to predict structure and 3D motion from real-world image pairs.
  arXiv  Detail & Related papers  (2024-12-12T18:59:54Z)
• Driv3R: Learning Dense 4D Reconstruction for Autonomous Driving [116.10577967146762]
  We propose Driv3R, a framework that directly regresses per-frame point maps from multi-view image sequences.<n>We employ a 4D flow predictor to identify moving objects within the scene to direct our network focus more on reconstructing these dynamic regions.<n>Driv3R outperforms previous frameworks in 4D dynamic scene reconstruction, achieving 15x faster inference speed.
  arXiv  Detail & Related papers  (2024-12-09T18:58:03Z)
• MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion [118.74385965694694]
  We present Motion DUSt3R (MonST3R), a novel geometry-first approach that directly estimates per-timestep geometry from dynamic scenes. By simply estimating a pointmap for each timestep, we can effectively adapt DUST3R's representation, previously only used for static scenes, to dynamic scenes. We show that by posing the problem as a fine-tuning task, identifying several suitable datasets, and strategically training the model on this limited data, we can surprisingly enable the model to handle dynamics.
  arXiv  Detail & Related papers  (2024-10-04T18:00:07Z)
• S4D: Streaming 4D Real-World Reconstruction with Gaussians and 3D Control Points [30.46796069720543]
  We introduce a novel approach for streaming 4D real-world reconstruction utilizing discrete 3D control points. This method physically models local rays and establishes a motion-decoupling coordinate system. By effectively merging traditional graphics with learnable pipelines, it provides a robust and efficient local 6-degrees-of-freedom (6 DoF) motion representation.
  arXiv  Detail & Related papers  (2024-08-23T12:51:49Z)
• Fast Encoder-Based 3D from Casual Videos via Point Track Processing [22.563073026889324]
  We present TracksTo4D, a learning-based approach that enables inferring 3D structure and camera positions from dynamic content originating from casual videos. TracksTo4D is trained in an unsupervised way on a dataset of casual videos. Experiments show that TracksTo4D can reconstruct a temporal point cloud and camera positions of the underlying video with accuracy comparable to state-of-the-art methods.
  arXiv  Detail & Related papers  (2024-04-10T15:37:00Z)
• DO3D: Self-supervised Learning of Decomposed Object-aware 3D Motion and Depth from Monocular Videos [76.01906393673897]
  We propose a self-supervised method to jointly learn 3D motion and depth from monocular videos. Our system contains a depth estimation module to predict depth, and a new decomposed object-wise 3D motion (DO3D) estimation module to predict ego-motion and 3D object motion. Our model delivers superior performance in all evaluated settings.
  arXiv  Detail & Related papers  (2024-03-09T12:22:46Z)
• Tensor4D : Efficient Neural 4D Decomposition for High-fidelity Dynamic Reconstruction and Rendering [31.928844354349117]
  We propose an efficient 4D tensor decomposition method for dynamic scenes. We show that our method is able to achieve high-quality dynamic reconstruction and rendering from sparse-view camera or even a monocular camera. The code and dataset will be released atliuyebin.com/tensor4d-tensor4d.html.
  arXiv  Detail & Related papers  (2022-11-21T16:04:45Z)
• 4D-Net for Learned Multi-Modal Alignment [87.58354992455891]
  We present 4D-Net, a 3D object detection approach, which utilizes 3D Point Cloud and RGB sensing information, both in time. We are able to incorporate the 4D information by performing a novel connection learning across various feature representations and levels of abstraction, as well as by observing geometric constraints.
  arXiv  Detail & Related papers  (2021-09-02T16:35:00Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.