在ImageNet图像上训练的卷积神经网络是否戴着玫瑰色眼镜?:ImageNet、计算机断层扫描、磁共振、胸部X线和床旁超声图像质量的定量比较
本研究采用盲/无参考图像空间质量评估器定量比较5500张ImageNet图像与2700张胸部X线、2300张CT、1800张磁共振和18000张床旁超声图像的质量。ImageNet图像质量评分为21.7,胸部X线为13.2,CT为35.1,磁共振为31.6,床旁超声为56.6,差异均具有统计学意义(P ≤ .000001)。结论提示超声图像质量与ImageNet及其他医学图像显著不同,可能影响卷积神经网络训练策略。
要点
- 床旁超声图像质量评分56.6,显著差于ImageNet和其他医学图像。
- ImageNet与床旁超声的图像质量差异最大(d=2.38)。
- 作者建议超声深度学习开发者考虑从零开始预训练网络。
核心结论
- 床旁超声图像质量与ImageNet及其他医学图像显著不同,对卷积神经网络训练有重要影响,超声图像可能更显著。
研究设计与样本
- 采用定量图像质量评估技术(盲/无参考图像空间质量评估器)比较像素复杂度、关系、变化和区分特征。
- 比较5500张ImageNet图像与2700张胸部X线、2300张CT、1800张磁共振和18000张床旁超声图像。
- 图像质量评分范围0至100(最差),采用单因素方差分析和标准化均数差效应量(d)。
诊断标准与关键数值
- ImageNet图像质量评分为21.7(95% CI, 0.41),胸部X线为13.2(95% CI, 0.28),CT为35.1(95% CI, 0.79),磁共振为31.6(95% CI, 0.75),床旁超声为56.6(95% CI, 0.21)。
- ImageNet与所有医学图像之间的差异均具有统计学意义(P ≤ .000001)。
- ImageNet与床旁超声的图像质量差异最大(d = 2.38)。
临床应用与阅片要点
- 超声深度学习开发者应考虑在超声图像上从零开始预训练网络。
- 用于CT、胸部X线和磁共振图像的训练技术可能不适用于超声。
局限性
- 原文未报告局限性。
以上解读依据论文摘要(全文未开放获取或许可不允许翻译转载)。
英文原文(摘要)· 点击展开
OBJECTIVES: Deep learning for medical imaging analysis uses convolutional neural networks pretrained on ImageNet (Stanford Vision Lab, Stanford, CA). Little is known about how such color- and scene-rich standard training images compare quantitatively to medical images. We sought to quantitatively compare ImageNet images to point-of-care ultrasound (POCUS), computed tomographic (CT), magnetic resonance (MR), and chest x-ray (CXR) images.
METHODS: Using a quantitative image quality assessment technique (Blind/Referenceless Image Spatial Quality Evaluator), we compared images based on pixel complexity, relationships, variation, and distinguishing features. We compared 5500 ImageNet images to 2700 CXR, 2300 CT, 1800 MR, and 18,000 POCUS images. Image quality results ranged from 0 to 100 (worst). A 1-way analysis of variance was performed, and the standardized mean-difference effect size value (d) was calculated.
RESULTS: ImageNet images showed the best image quality rating of 21.7 (95% confidence interval [CI], 0.41) except for CXR at 13.2 (95% CI, 0.28), followed by CT at 35.1 (95% CI, 0.79), MR at 31.6 (95% CI, 0.75), and POCUS at 56.6 (95% CI, 0.21). The differences between ImageNet and all of the medical images were statistically significant (P ≤ .000001). The greatest difference in image quality was between ImageNet and POCUS (d = 2.38).
CONCLUSIONS: Point-of-care ultrasound (US) quality is significantly different from that of ImageNet and other medical images. This brings considerable implications for convolutional neural network training with medical images for various applications, which may be even more significant in the case of US images. Ultrasound deep learning developers should consider pretraining networks from scratch on US images, as training techniques used for CT, CXR, and MR images may not apply to US.
引用
- Blaivas L, Blaivas M. Are Convolutional Neural Networks Trained on ImageNet Images Wearing Rose-Colored Glasses?: A Quantitative Comparison of ImageNet, Computed Tomographic, Magnetic Resonance, Chest X-Ray, and Point-of-Care Ultrasound Images for Quality. J Ultrasound Med. 2020. PMID: 32757235. doi:10.1002/jum.15413
供执业医师工作参考,不替代临床判断;数值与分级以指南原文为准。