arXiv cs.CL· Yuanhe Tian, Yan Song·· 11 小时前AI 评分31
SliceWorld：面向 CT 报告生成的预测性可控世界状态模型
SliceWorld: A Predictive and Controllable World-State Model for CT Report Generation
AI 导读
SliceWorld 是一个面向 CT 报告生成的 CT 专用世界状态框架，将轴位 CT 扫描视为沿 z 轴的有序序列，把前缀 CT 证据编码为含解剖、病灶与不确定性成分的因子感知潜在状态，并投影为 world tokens 用于多步未来切片特征预测、病灶因子干预和基于 LLM 的报告生成。
正文 · 原文
Abstract:CT report generation (CTRG) requires models to summarize three-dimensional anatomical context and pathological findings from hundreds of axial slices. Existing methods typically learn a direct image-to-text mapping, providing limited mechanisms for modeling how CT evidence evolves across slices or how reports respond to controlled changes in latent lesion-related factors. We propose SliceWorld, a CT-specific world-state framework that treats an axial CT scan as an ordered sequence along the z-axis. SliceWorld encodes prefix CT evidence into factor-aware latent states containing anatomy, lesion, and uncertainty components, and projects these states into world tokens used for multi-step future-slice feature prediction, lesion-factor intervention, and LLM-based report generation. The model is first pretrained on CT slice sequences with predictive, factor-aware, and counterfactual objectives, and is then fine-tuned on paired CT-report data. Experiments on M3D-Cap and CT-RATE show that SliceWorld improves natural language generation metrics and clinically oriented automatic evaluation. Further analyses demonstrate multi-horizon future-slice prediction, measurable factor alignment, reduced-slice robustness, and selective lesion-sensitive report modulation. Code is available at this https URL.
| Comments: | Accepted by NeurIPS 2026, 20 pages, 5 figures |
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Computation and Language (cs.CL) |
| Cite as: | arXiv:2605.24371 [cs.CV] |
| (or arXiv:2605.24371v2 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2605.24371 arXiv-issued DOI via DataCite |
|
| Journal reference: | NeurIPS 2026 |
Submission history
From: Yuanhe Tian [view email]
[v1]
Sat, 23 May 2026 03:18:56 UTC (2,185 KB)
[v2]
Fri, 6 Nov 2026 14:01:41 UTC (2,186 KB)
来源：arXiv cs.CL · arxiv.org