面向临床相关医学图像字幕生成的区域注意力增强Swin Transformer
Regional Attention-Enhanced Swin Transformer for Clinically Relevant Medical Image Captioning
AI总结：
该研究提出搭载轻量区域注意力模块的Swin-BART医学图像字幕系统，在ROCO数据集上多项指标达最优，兼具紧凑性与可解释性，支持有人参与的安全临床研究应用。
摘要：
自动化医学图像字幕生成可将复杂的放射学图像转换为诊断性叙述，为报告工作流提供支持。我们提出了一种搭载轻量级区域注意力模块的Swin-BART编码器-解码器系统，该模块会在交叉注意力之前放大诊断关键区域。我们在ROCO（放射学图像字幕数据集）上对模型进行训练和评估，其在保持紧凑性与可解释性的同时，实现了当前最优的语义保真度。我们报告的结果为三次随机种子实验的均值±标准差，并包含95%置信区间。与基线模型相比，我们的方法提升了ROUGE得分（所提方法0.603，ResNet-CNN 0.356，BLIP2-OPT 0.255）和BERTScore（所提方法0.807，BLIP2-OPT 0.645，ResNet-CNN 0.623），且BLEU、CIDEr和METEOR得分具备竞争力。我们进一步提供了消融实验（区域注意力开启/关闭及词元数量扫描）、跨模态分析（CT/MRI/X射线）、配对显著性检验，以及可视化驱动每条描述的区域的定性热力图。解码采用束搜索（束大小=4）、长度惩罚=1.1、no_repeat_ngram_size=3、最大长度=128。所提设计可生成准确、符合临床表述的字幕并提供透明的区域归因，支持有人参与的安全研究应用。
英文摘要：
Automated medical image captioning translates complex radiological images into diagnostic narratives that can support reporting workflows. We present a Swin-BART encoder-decoder system with a lightweight regional attention module that amplifies diagnostically salient regions before cross-attention. Trained and evaluated on ROCO, our model achieves state-of-the-art semantic fidelity while remaining compact and interpretable. We report results as mean$\pm$std over three seeds and include $95\%$ confidence intervals. Compared with baselines, our approach improves ROUGE (proposed 0.603, ResNet-CNN 0.356, BLIP2-OPT 0.255) and BERTScore (proposed 0.807, BLIP2-OPT 0.645, ResNet-CNN 0.623), with competitive BLEU, CIDEr, and METEOR. We further provide ablations (regional attention on/off and token-count sweep), per-modality analysis (CT/MRI/X-ray), paired significance tests, and qualitative heatmaps that visualize the regions driving each description. Decoding uses beam search (beam size $=4$), length penalty $=1.1$, $no\_repeat\_ngram\_size$ $=3$, and max length $=128$. The proposed design yields accurate, clinically phrased captions and transparent regional attributions, supporting safe research use with a human in the loop.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
