Research map / 3D vision / Depth and stereo
Easi3R: Estimating Disentangled Motion from DUSt3R Without Training
Xingyu Chen, Yue Chen, Yuliang Xiu, Andreas Geiger, Anpei Chen
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add A_ref mean attention map to segmentation (ablation) | Easi3R segmentation · DAVIS-16 (w/o SAM2, DUSt3R backbone) | JM: 53.1 -> 33.3 when removed | Improved |
| add A_src mean attention map (ablation) | Easi3R segmentation · DAVIS-16 (w/o SAM2, DUSt3R backbone) | JM: 53.1 -> 45.1 when removed | Improved |
| add A_src std attention map (ablation) | Easi3R segmentation · DAVIS-16 (w/o SAM2, DUSt3R backbone) | JM: 53.1 -> 42.3 when removed | Improved |
| add A_ref std attention map (ablation) | Easi3R segmentation · DAVIS-16 (w/o SAM2, DUSt3R backbone) | JM: 53.1 -> 47.7 when removed | Improved |
| add cross-frame feature clustering (ablation) | Easi3R segmentation · DAVIS-16 (w/o SAM2, DUSt3R backbone) | JM: 53.1 -> 40.0 when removed | Improved |
| use window size 3 instead of 5 (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 76.0 | Improved |
| use window size 7 instead of 5 (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 66.9 | Hurt |
| use 16 clusters instead of 64 (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 67.4 | Hurt |
| use 32 clusters instead of 64 (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 71.6 | Improved |
| use fixed threshold 0.5 instead of Otsu (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 61.6 | Hurt |
| use fixed threshold 0.7 instead of Otsu (ablation) | Easi3R (MonST3R backbone) · DAVIS-16 (w/ SAM2) | JM: 70.7 -> 70.2 | No clear effect |
| re-weight only the reference decoder instead of both branches (ablation) | Easi3R (DUSt3R backbone) · DyCheck pose | ATE: 0.030 -> 0.029 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- MonST3R — MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion (ICLR 2025) · better on 5 of 5 shared results
- DUSt3R — DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024) · better on 5 of 5 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- MonST3R: A Simple Approach for Estimating Geometry in the Presence of Motion (ICLR 2025 · similarity 0.89)
- 4D3R: Motion-Aware Neural Reconstruction and Rendering of Dynamic Scenes from Monocular Videos (NeurIPS 2025 · similarity 0.89)
- Stereo4D: Learning How Things Move in 3D from Internet Stereo Videos (CVPR 2025 · similarity 0.89)
- MV-DUSt3R+: Single-Stage Scene Reconstruction from Sparse Views In 2 Seconds (CVPR 2025 · similarity 0.88)
- DUSt3R: Geometric 3D Vision Made Easy (CVPR 2024 · similarity 0.88)
- Flux4D: Flow-based Unsupervised 4D Reconstruction (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
