利用条件StyleGAN和隐空间操纵实现可控视网膜图像合成以改善糖尿病视网膜病变的诊断与分级
Controllable retinal image synthesis using conditional StyleGAN and latent space manipulation for improved diagnosis and grading of diabetic retinopathy
AI总结：
针对糖尿病视网膜病变分级标注数据不足的问题，提出基于条件StyleGAN与SeFa隐空间操纵的可控图像合成及增强策略，在APTOS 2019数据集上提升了检测与分级模型性能。
摘要：
糖尿病视网膜病变（DR）是糖尿病的并发症，以视网膜组织内的血管损伤为特征。及时检测对于降低视力丧失风险至关重要。然而，标注数据的短缺（尤其是重症病例的标注数据）阻碍了鲁棒分级模型的训练。本文提出了一种可控生成高保真、多样化DR眼底图像的框架，以此提升DR分级与检测任务中分类器的性能。我们仅通过条件StyleGAN（条件风格生成对抗网络）就实现了对生成图像中DR严重程度及视觉特征（视盘、血管结构、病变区域）的全面控制，无需特征掩码或辅助网络。具体而言，我们利用SeFa（隐空间语义分解）算法识别隐空间内有意义的语义，对按等级条件生成的DR图像进行操纵，进一步增强数据集多样性。此外，我们提出了一种新颖有效的基于SeFa的数据增强策略，帮助分类器聚焦于判别性区域同时忽略冗余特征。采用该方法，用于DR检测的ResNet50模型取得了98.09%的准确率、99.44%的特异性、99.45%的精确率以及98.09%的F1分数。此外，将条件StyleGAN生成的合成图像纳入ResNet50的DR分级训练中，取得了83.33%的准确率、87.64%的二次加权卡帕分数、95.67%的特异性以及72.24%的精确率。在APTOS 2019数据集上开展的大量实验证明，生成图像具有极高的真实感，且我们的分类器性能优于近期相关研究。
英文摘要：
Diabetic retinopathy (DR) is a consequence of diabetes mellitus characterized by vascular damage within the retinal tissue. Timely detection is paramount to mitigate the risk of vision loss. However, training robust grading models is hindered by a shortage of annotated data, particularly for severe cases. This paper proposes a framework for controllably generating high-fidelity and diverse DR fundus images, thereby improving classifier performance in DR grading and detection. We achieve comprehensive control over DR severity and visual features (optic disc, vessel structure, lesion areas) within generated images solely through a conditional StyleGAN, eliminating the need for feature masks or auxiliary networks. Specifically, leveraging the SeFa algorithm to identify meaningful semantics within the latent space, we manipulate the DR images generated conditionally on grades, further enhancing the dataset diversity. Additionally, we propose a novel, effective SeFa-based data augmentation strategy, helping the classifier focus on discriminative regions while ignoring redundant features. Using this approach, a ResNet50 model trained for DR detection achieves 98.09% accuracy, 99.44% specificity, 99.45% precision, and an F1-score of 98.09%. Moreover, incorporating synthetic images generated by conditional StyleGAN into ResNet50 training for DR grading yields 83.33% accuracy, a quadratic kappa score of 87.64%, 95.67% specificity, and 72.24% precision. Extensive experiments conducted on the APTOS 2019 dataset demonstrate the exceptional realism of the generated images and the superior performance of our classifier compared to recent studies.
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。
