DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method
Fuente:
arXiv
Salvato in:
| Autori principali: | , , , , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2025
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866912781406568448 |
|---|---|
| author | Zhang, Qingwen Zhu, Xiaomeng Zhang, Yushan Cai, Yixi Andersson, Olov Jensfelt, Patric |
| author_facet | Zhang, Qingwen Zhu, Xiaomeng Zhang, Yushan Cai, Yixi Andersson, Olov Jensfelt, Patric |
| contents | Previous dominant methods for scene flow estimation focus mainly on input from two consecutive frames, neglecting valuable information in the temporal domain. While recent trends shift towards multi-frame reasoning, they suffer from rapidly escalating computational costs as the number of frames grows. To leverage temporal information more efficiently, we propose DeltaFlow ($Δ$Flow), a lightweight 3D framework that captures motion cues via a $Δ$ scheme, extracting temporal features with minimal computational cost, regardless of the number of frames. Additionally, scene flow estimation faces challenges such as imbalanced object class distributions and motion inconsistency. To tackle these issues, we introduce a Category-Balanced Loss to enhance learning across underrepresented classes and an Instance Consistency Loss to enforce coherent object motion, improving flow accuracy. Extensive evaluations on the Argoverse 2, Waymo and nuScenes datasets show that $Δ$Flow achieves state-of-the-art performance with up to 22% lower error and $2\times$ faster inference compared to the next-best multi-frame supervised method, while also demonstrating a strong cross-domain generalization ability. The code is open-sourced at https://github.com/Kin-Zhang/DeltaFlow along with trained model weights. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2508_17054 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method Zhang, Qingwen Zhu, Xiaomeng Zhang, Yushan Cai, Yixi Andersson, Olov Jensfelt, Patric Computer Vision and Pattern Recognition Robotics Previous dominant methods for scene flow estimation focus mainly on input from two consecutive frames, neglecting valuable information in the temporal domain. While recent trends shift towards multi-frame reasoning, they suffer from rapidly escalating computational costs as the number of frames grows. To leverage temporal information more efficiently, we propose DeltaFlow ($Δ$Flow), a lightweight 3D framework that captures motion cues via a $Δ$ scheme, extracting temporal features with minimal computational cost, regardless of the number of frames. Additionally, scene flow estimation faces challenges such as imbalanced object class distributions and motion inconsistency. To tackle these issues, we introduce a Category-Balanced Loss to enhance learning across underrepresented classes and an Instance Consistency Loss to enforce coherent object motion, improving flow accuracy. Extensive evaluations on the Argoverse 2, Waymo and nuScenes datasets show that $Δ$Flow achieves state-of-the-art performance with up to 22% lower error and $2\times$ faster inference compared to the next-best multi-frame supervised method, while also demonstrating a strong cross-domain generalization ability. The code is open-sourced at https://github.com/Kin-Zhang/DeltaFlow along with trained model weights. |
| title | DeltaFlow: An Efficient Multi-frame Scene Flow Estimation Method |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2508.17054 |