用扩散模型生成高分辨率胎儿超声图像，提升下游分类性能。
A Foundational EDM2-Based Generative Model for High-Resolution Synthetic Fetal Ultrasound Imaging from Open Datasets
- 基于EDM2架构，融合多个公开数据集训练生成512x512图像。
- 微调后胎儿切面分类准确率达93.36%，优于仅用真实数据训练。
- 生成图像临床评估真实感均分2.67/5，适合数据稀缺场景研究。
产前超声成像是评估胎儿健康的关键手段，但人工智能进展受限于数据稀缺、隐私限制及难以标注。本文提出一种基于EDM2扩散架构的高分辨率胎儿超声合成框架，利用多个公开数据集训练，生成512×512像素、涵盖六类解剖结构的图像。该方法在图像质量上表现更优（更低的FID分数），并显著提升下游胎儿切面分类性能，微调后集成准确率达93.36%，超越仅使用真实数据训练的基线。由资深胎儿超声医师（10年以上经验）对100张图像进行临床评估，生成图像平均真实感得分为2.67/5，真实图像评分更高。常见伪影包括平滑化、斑点不规则及解剖不一致。代码、数据、模型及其他资源可在https://github.com/xfetus/fetal-ultrasound-edm2获取。
原文摘要 · Abstract (English)
Prenatal ultrasound imaging is key for assessing fetal health, but AI progress is limited by scarce, privacy-restricted, and hard-to-annotate datasets. We propose a high-resolution fetal ultrasound synthesis framework based on the EDM2 diffusion architecture, trained on multiple public datasets to generate 512x512 images across six anatomical classes. Our method achieved improved image quality with lower FID scores and enhanced downstream fetal plane classification, reaching 93.36% ensemble accuracy after fine-tuning, surpassing real-data-only training. Clinical evaluation by an experienced fetal ultrasound specialist (10+ years) on 100 images yielded a mean realism score of 2.67/5, with real images rated higher than synthetic. Artefacts included smoothing, speckle irregularities, and anatomical inconsistencies. Code, data, models and other resources to reproduce this work are available at https://github.com/xfetus/fetal-ultrasound-edm2.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。