WildRoadBench: A Wild Aerial Road-Damage Grounding Benchmark for Vision-Language Models and Autonomous Agents
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_ | 1866917512166244352 |
|---|---|
| author | Liu, Bingnan Cui, Chenhang Huang, Rui Luo, Jiani Shen, Zhirong Wang, Tinghao Huang, Xiande Meng, Lingbei Shen, Fei Zhang, An |
| author_facet | Liu, Bingnan Cui, Chenhang Huang, Rui Luo, Jiani Shen, Zhirong Wang, Tinghao Huang, Xiande Meng, Lingbei Shen, Fei Zhang, An |
| contents | We introduce WildRoadBench, a wild aerial road-damage grounding benchmark that couples direct visual grounding by vision-language models with autonomous research-and-engineering by LLM-driven agents on a single professionally annotated UAV corpus. The same image set and the same per-class AP_50 metric are evaluated under two protocols. The VLM Track measures whether a fixed VLM can localise domain-specific damage from one image and one short prompt under a unified prompting, decoding and parsing pipeline. The Agent Track measures whether an autonomous agent, given only a written task brief, a small exploratory slice and a fixed interaction budget, can search the public web, adapt pretrained components, write training and inference code, and submit predictions through a scalar-feedback oracle on a hidden holdout. We benchmark a broad pool of closed-source frontier models and open-source VLMs together with several frontier LLM-driven agents. Both routes remain far from reliable performance in this wild setting: closed-source frontier models lead the VLM leaderboard but still leave more than half of the metric on the table; open-source grounders plateau well below them, and newer generations or reasoning-style variants do not consistently improve grounding; small targets collapse for every open-source model; agents lag the strongest VLM despite richer affordances, and several fail to land a valid submission within the budget. We release the code and data at https://anonymous.4open.science/r/wildroadbench-0607 to support reproducible follow-up research. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_20306 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | WildRoadBench: A Wild Aerial Road-Damage Grounding Benchmark for Vision-Language Models and Autonomous Agents Liu, Bingnan Cui, Chenhang Huang, Rui Luo, Jiani Shen, Zhirong Wang, Tinghao Huang, Xiande Meng, Lingbei Shen, Fei Zhang, An Computer Vision and Pattern Recognition Machine Learning I.2.10; I.4.8; I.2.7; I.2.6 We introduce WildRoadBench, a wild aerial road-damage grounding benchmark that couples direct visual grounding by vision-language models with autonomous research-and-engineering by LLM-driven agents on a single professionally annotated UAV corpus. The same image set and the same per-class AP_50 metric are evaluated under two protocols. The VLM Track measures whether a fixed VLM can localise domain-specific damage from one image and one short prompt under a unified prompting, decoding and parsing pipeline. The Agent Track measures whether an autonomous agent, given only a written task brief, a small exploratory slice and a fixed interaction budget, can search the public web, adapt pretrained components, write training and inference code, and submit predictions through a scalar-feedback oracle on a hidden holdout. We benchmark a broad pool of closed-source frontier models and open-source VLMs together with several frontier LLM-driven agents. Both routes remain far from reliable performance in this wild setting: closed-source frontier models lead the VLM leaderboard but still leave more than half of the metric on the table; open-source grounders plateau well below them, and newer generations or reasoning-style variants do not consistently improve grounding; small targets collapse for every open-source model; agents lag the strongest VLM despite richer affordances, and several fail to land a valid submission within the budget. We release the code and data at https://anonymous.4open.science/r/wildroadbench-0607 to support reproducible follow-up research. |
| title | WildRoadBench: A Wild Aerial Road-Damage Grounding Benchmark for Vision-Language Models and Autonomous Agents |
| topic | Computer Vision and Pattern Recognition Machine Learning I.2.10; I.4.8; I.2.7; I.2.6 |
| url | https://arxiv.org/abs/2605.20306 |