Schaudt Daniel, Späte Christian, von Schwerin Reinhold, Reichert Manfred, von Schwerin Marianne, Beer Meinrad, Kloth Christopher
Institute of Databases and Information Systems, Ulm University, James-Franck-Ring, 89081 Ulm, Germany.
DASU Transferzentrum für Digitalisierung, Analytics und Data Science Ulm, Olgastraße 94, 89073 Ulm, Germany.
Bioengineering (Basel). 2023 Dec 14;10(12):1421. doi: 10.3390/bioengineering10121421.
In medical imaging, deep learning models serve as invaluable tools for expediting diagnoses and aiding specialized medical professionals in making clinical decisions. However, effectively training deep learning models typically necessitates substantial quantities of high-quality data, a resource often lacking in numerous medical imaging scenarios. One way to overcome this deficiency is to artificially generate such images. Therefore, in this comparative study we train five generative models to artificially increase the amount of available data in such a scenario. This synthetic data approach is evaluated on a a downstream classification task, predicting four causes for pneumonia as well as healthy cases on 1082 chest X-ray images. Quantitative and medical assessments show that a Generative Adversarial Network (GAN)-based approach significantly outperforms more recent diffusion-based approaches on this limited dataset with better image quality and pathological plausibility. We show that better image quality surprisingly does not translate to improved classification performance by evaluating five different classification models and varying the amount of additional training data. Class-specific metrics like precision, recall, and F1-score show a substantial improvement by using synthetic images, emphasizing the data rebalancing effect of less frequent classes. However, overall performance does not improve for most models and configurations, except for a DreamBooth approach which shows a +0.52 improvement in overall accuracy. The large variance of performance impact in this study suggests a careful consideration of utilizing generative models for limited data scenarios, especially with an unexpected negative correlation between image quality and downstream classification improvement.
在医学成像中，深度学习模型是加快诊断速度并协助专业医疗人员做出临床决策的宝贵工具。然而，有效训练深度学习模型通常需要大量高质量数据，而在众多医学成像场景中，这种资源往往匮乏。克服这一不足的一种方法是人工生成此类图像。因此，在这项对比研究中，我们训练了五个生成模型，以便在这种场景下人工增加可用数据量。这种合成数据方法在一个下游分类任务中进行评估，该任务是在1082张胸部X光图像上预测肺炎的四种病因以及健康病例。定量和医学评估表明，在这个有限的数据集上，基于生成对抗网络（GAN）的方法明显优于最近基于扩散的方法，生成的图像质量更高，病理合理性更强。通过评估五种不同的分类模型并改变额外训练数据的数量，我们发现令人惊讶的是，更好的图像质量并没有转化为更高的分类性能。像精确率、召回率和F1分数这样的特定类别指标显示，使用合成图像有显著改善，这突出了较少出现的类别的数据重新平衡效果。然而，除了一种DreamBooth方法在整体准确率上提高了0.52之外，大多数模型和配置的整体性能并没有提高。本研究中性能影响的巨大差异表明，在有限数据场景中使用生成模型时需要谨慎考虑，特别是在图像质量与下游分类改进之间存在意外负相关的情况下。
