Is Your Driving World Model an All-Around Player?
Fuente:
arXiv
Guardado en:
| Autores principales: | , , , , , , , , , , , , , , , , , , , , , , |
|---|---|
| Formato: | Preprint |
| Publicado: |
2026
|
| Materias: | |
| Acceso en línea: | |
| Etiquetas: |
Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
|
| _version_ | 1866910209513881600 |
|---|---|
| author | Kong, Lingdong Liang, Ao Yan, Tianyi Liu, Hongsi Yang, Wesley Huang, Ziqi Sun, Xian Yin, Wei Zuo, Jialong Hu, Yixuan Zhu, Dekai Lu, Dongyue Liu, Youquan Jiang, Guangfeng Li, Linfeng Li, Xiangtai Zhuo, Long Ng, Lai Xing Cottereau, Benoit R. Gao, Changxin Pan, Liang Ooi, Wei Tsang Liu, Ziwei |
| author_facet | Kong, Lingdong Liang, Ao Yan, Tianyi Liu, Hongsi Yang, Wesley Huang, Ziqi Sun, Xian Yin, Wei Zuo, Jialong Hu, Yixuan Zhu, Dekai Lu, Dongyue Liu, Youquan Jiang, Guangfeng Li, Linfeng Li, Xiangtai Zhuo, Long Ng, Lai Xing Cottereau, Benoit R. Gao, Changxin Pan, Liang Ooi, Wei Tsang Liu, Ziwei |
| contents | Today's driving world models can generate remarkably realistic dash-cam videos, yet no single model excels universally. Some generate photorealistic textures but violate basic physics; others maintain geometric consistency but fail when subjected to closed-loop planning. This disconnect exposes a critical gap: the field evaluates how real generated worlds appear, but rarely whether they behave realistically. We introduce WorldLens, a unified benchmark that measures world-model fidelity across the full spectrum, from pixel quality and 4D geometry to closed-loop driving and human perceptual alignment, through five complementary aspects and 24 standardized dimensions. Our evaluation of six representative models reveals that no existing approach dominates across all axes: texture-rich models violate geometry, geometry-aware models lack behavioral fidelity, and even the strongest performers achieve only 2-3 out of 10 on human realism ratings. To bridge algorithmic metrics with human perception, we further contribute WorldLens-26K, a 26,808-entry human-annotated preference dataset pairing numerical scores with textual rationales, and WorldLens-Agent, a vision-language evaluator distilled from these judgments that enables scalable, explainable auto-assessment. Together, the benchmark, dataset, and agent form a unified ecosystem for assessing generated worlds not merely by visual appeal, but by physical and behavioral fidelity. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_10858 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Is Your Driving World Model an All-Around Player? Kong, Lingdong Liang, Ao Yan, Tianyi Liu, Hongsi Yang, Wesley Huang, Ziqi Sun, Xian Yin, Wei Zuo, Jialong Hu, Yixuan Zhu, Dekai Lu, Dongyue Liu, Youquan Jiang, Guangfeng Li, Linfeng Li, Xiangtai Zhuo, Long Ng, Lai Xing Cottereau, Benoit R. Gao, Changxin Pan, Liang Ooi, Wei Tsang Liu, Ziwei Computer Vision and Pattern Recognition Robotics Today's driving world models can generate remarkably realistic dash-cam videos, yet no single model excels universally. Some generate photorealistic textures but violate basic physics; others maintain geometric consistency but fail when subjected to closed-loop planning. This disconnect exposes a critical gap: the field evaluates how real generated worlds appear, but rarely whether they behave realistically. We introduce WorldLens, a unified benchmark that measures world-model fidelity across the full spectrum, from pixel quality and 4D geometry to closed-loop driving and human perceptual alignment, through five complementary aspects and 24 standardized dimensions. Our evaluation of six representative models reveals that no existing approach dominates across all axes: texture-rich models violate geometry, geometry-aware models lack behavioral fidelity, and even the strongest performers achieve only 2-3 out of 10 on human realism ratings. To bridge algorithmic metrics with human perception, we further contribute WorldLens-26K, a 26,808-entry human-annotated preference dataset pairing numerical scores with textual rationales, and WorldLens-Agent, a vision-language evaluator distilled from these judgments that enables scalable, explainable auto-assessment. Together, the benchmark, dataset, and agent form a unified ecosystem for assessing generated worlds not merely by visual appeal, but by physical and behavioral fidelity. |
| title | Is Your Driving World Model an All-Around Player? |
| topic | Computer Vision and Pattern Recognition Robotics |
| url | https://arxiv.org/abs/2605.10858 |