GA-Drive: Geometry-Appearance Decoupled Modeling for Free-viewpoint Driving Scene Generation
Fuente:
arXiv
Enregistré dans:
| Auteurs principaux: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Publié: |
2026
|
| Sujets: | |
| Accès en ligne: | |
| Tags: |
Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
|
| _version_ | 1866912964192239616 |
|---|---|
| author | Zhang, Hao Fan, Lue Wang, Qitai Li, Wenbo Wu, Zehuan Lu, Lewei Zhang, Zhaoxiang Li, Hongsheng |
| author_facet | Zhang, Hao Fan, Lue Wang, Qitai Li, Wenbo Wu, Zehuan Lu, Lewei Zhang, Zhaoxiang Li, Hongsheng |
| contents | A free-viewpoint, editable, and high-fidelity driving simulator is crucial for training and evaluating end-to-end autonomous driving systems. In this paper, we present GA-Drive, a novel simulation framework capable of generating camera views along user-specified novel trajectories through Geometry-Appearance Decoupling and Diffusion-Based Generation. Given a set of images captured along a recorded trajectory and the corresponding scene geometry, GA-Drive synthesizes novel pseudo-views using geometry information. These pseudo-views are then transformed into photorealistic views using a trained video diffusion model. In this way, we decouple the geometry and appearance of scenes. An advantage of such decoupling is its support for appearance editing via state-of-the-art video-to-video editing techniques, while preserving the underlying geometry, enabling consistent edits across both original and novel trajectories. Extensive experiments demonstrate that GA-Drive substantially outperforms existing methods in terms of NTA-IoU, NTL-IoU, and FID scores. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_20673 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | GA-Drive: Geometry-Appearance Decoupled Modeling for Free-viewpoint Driving Scene Generation Zhang, Hao Fan, Lue Wang, Qitai Li, Wenbo Wu, Zehuan Lu, Lewei Zhang, Zhaoxiang Li, Hongsheng Computer Vision and Pattern Recognition A free-viewpoint, editable, and high-fidelity driving simulator is crucial for training and evaluating end-to-end autonomous driving systems. In this paper, we present GA-Drive, a novel simulation framework capable of generating camera views along user-specified novel trajectories through Geometry-Appearance Decoupling and Diffusion-Based Generation. Given a set of images captured along a recorded trajectory and the corresponding scene geometry, GA-Drive synthesizes novel pseudo-views using geometry information. These pseudo-views are then transformed into photorealistic views using a trained video diffusion model. In this way, we decouple the geometry and appearance of scenes. An advantage of such decoupling is its support for appearance editing via state-of-the-art video-to-video editing techniques, while preserving the underlying geometry, enabling consistent edits across both original and novel trajectories. Extensive experiments demonstrate that GA-Drive substantially outperforms existing methods in terms of NTA-IoU, NTL-IoU, and FID scores. |
| title | GA-Drive: Geometry-Appearance Decoupled Modeling for Free-viewpoint Driving Scene Generation |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2602.20673 |