Thought Graph Traversal for Test-time Scaling in Chest X-ray VLLMs
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866913078420963328 |
|---|---|
| author | Yao, Yue Wen, Zelin Tong, Yan Tian, Xinyu Li, Xuqing Ma, Xiao Xu, Dongliang Gedeon, Tom |
| author_facet | Yao, Yue Wen, Zelin Tong, Yan Tian, Xinyu Li, Xuqing Ma, Xiao Xu, Dongliang Gedeon, Tom |
| contents | Test-time scaling offers a promising way to improve the reasoning performance of vision-language large models (VLLMs) without additional training. In this paper, we explore a simple but effective approach for applying test-time scaling to chest X-ray report generation. Specifically, we introduce a lightweight Thought Graph Traversal (TGT) framework that guides the model to reason through organ-specific findings in a medically coherent order. This framework integrates structured medical priors into the prompt, enabling deeper and more logical analysis with no changes to the underlying model. To further enhance reasoning depth, we apply a reasoning budget forcing strategy that adjusts the model's inference depth at test time by dynamically extending its generation process. This simple yet powerful combination allows a frozen radiology VLLM to self-correct and generate more accurate, consistent chest X-ray reports. Our method outperforms baseline prompting approaches on standard benchmarks, and also reveals dataset biases through traceable reasoning paths. Code and prompts are open-sourced for reproducibility at https://github.com/glerium/Thought-Graph-Traversal |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_11989 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Thought Graph Traversal for Test-time Scaling in Chest X-ray VLLMs Yao, Yue Wen, Zelin Tong, Yan Tian, Xinyu Li, Xuqing Ma, Xiao Xu, Dongliang Gedeon, Tom Computer Vision and Pattern Recognition Test-time scaling offers a promising way to improve the reasoning performance of vision-language large models (VLLMs) without additional training. In this paper, we explore a simple but effective approach for applying test-time scaling to chest X-ray report generation. Specifically, we introduce a lightweight Thought Graph Traversal (TGT) framework that guides the model to reason through organ-specific findings in a medically coherent order. This framework integrates structured medical priors into the prompt, enabling deeper and more logical analysis with no changes to the underlying model. To further enhance reasoning depth, we apply a reasoning budget forcing strategy that adjusts the model's inference depth at test time by dynamically extending its generation process. This simple yet powerful combination allows a frozen radiology VLLM to self-correct and generate more accurate, consistent chest X-ray reports. Our method outperforms baseline prompting approaches on standard benchmarks, and also reveals dataset biases through traceable reasoning paths. Code and prompts are open-sourced for reproducibility at https://github.com/glerium/Thought-Graph-Traversal |
| title | Thought Graph Traversal for Test-time Scaling in Chest X-ray VLLMs |
| topic | Computer Vision and Pattern Recognition |
| url | https://arxiv.org/abs/2506.11989 |