# Clinically Structured Surrogate Rewards for Post-SFT Medical Image Captioning

Preprint · Hyun Jun Kim, Heeseung Shin, Changwon Lim

- Archive id: arxiv:2608.18654
- Published: 2026-08-19
- arXiv: https://arxiv.org/abs/2608.18654
- Citations: 0
- Source URL: https://arxiv.org/abs/2608.18654
- PDF: https://arxiv.org/pdf/2608.18654
- Provenance: arxiv

## Abstract

Medical image captioning requires translating heterogeneous visual evidence into concise clinical descriptions, where errors in findings, assertion states, or anatomical relations can alter clinical meaning despite surface-level fluency. Sequence-level policy optimization can directly optimize complete captions, but common rewards rely on global text similarity, direct image-caption compatibility, or unordered concept overlap, leaving visual neighborhoods and clinical-claim structure implicit. We propose a clinically structured surrogate reward framework for post-SFT medical image captioning. The framework combines biomedical semantic and short-range lexical fidelity with two structured rewards: distributional image-neighborhood alignment, which matches the medical-image-bank distributions induced by reference and generated captions, and clinical graph consistency, which applies maximum-weight one-to-one matching to entities, assertion states, and typed relations. The four rewards are independently normalized within each rollout group, combined with fixed relative weights, and optimized with GDPO. Across organizer-evaluated hidden test sets for the Standard and Synthetical ImageCLEFmedical Caption tracks and three vision-language backbones, the method improves Overall, Relevance, and Factuality over matched SFT baselines in all six backbone-track combinations, with average relative gains of 3.4%, 2.1%, and 5.8%, respectively. Ablations and paired diagnostics indicate that the structured rewards provide complementary signals, reducing image-neighborhood divergence and improving entity-assertion-relation consistency.

_Abstract reproduced under CC0-1.0 (arXiv descriptive metadata)._

Topics: [AI in Clinical Medicine](https://space-kitty.com/topics/ai-medicine)

---
Space Kitty Archive · https://space-kitty.com · API: https://space-kitty.com/api/v1 · MCP: https://space-kitty.com/mcp · Premium: $0.20/request via x402 (Base Sepolia (testnet)). Metadata: arXiv (CC0), OpenAlex (CC0), Europe PMC (CC BY/CC0 works), ClinicalTrials.gov.