Adaptive adversarial neural networks for the analysis of lossy and domain-shifted datasets of medical images.
Adaptive adversarial neural networks for the analysis of lossy and domain-shifted datasets of medical images.
复制标题
自适应对抗神经网络用于分析医学图像的有损和域偏移数据集。
DOI：
10.1038/s41551-021-00733-w
复制
发表时间：
2021-06
影响因子：
28.1
通讯作者：
Shafiee H
中科院分区：
文献类型：
--
作者：
Kanakasabapathy MK;Thirumalaraju P;Kandula H;Doshi F;Sivakumar AD;Kartik D;Gupta R;Pooniwala R;Branda JA;Tsibris AM;Kuritzkes DR;Petrozza JC;Bormann CL;Shafiee H
在用于基于图像的医疗诊断的机器学习中，监督卷积神经网络通常使用使用高分辨率成像系统获得的大型专业注释数据集进行训练。此外，当应用于具有不同分布的数据集时，网络的性能可能会大幅下降。在这里，我们展示了对抗学习可以用来开发高性能的网络，这些网络在不同图像质量的未注释医学图像上训练。具体来说，我们使用廉价的便携式光学系统获得的低质量图像来训练网络，用于评估人类胚胎，人类精子形态的量化和血液中疟疾感染的诊断，并表明网络在不同的数据分布中表现良好。我们还表明，对抗性学习可以用于未标记的数据从看不见的域转移数据集，以适应预训练的监督网络新的分布，即使从原始分布的数据是不可用的。自适应对抗网络可以扩展经验证的神经网络模型的使用，以评估从不同质量的多个成像系统收集的数据，而不会损害存储在网络中的知识。
In machine learning for image-based medical diagnostics, supervised convolutional neural networks are typically trained with large and expertly annotated datasets obtained using high-resolution imaging systems. Moreover, the network’s performance can degrade substantially when applied to a dataset with a different distribution. Here, we show that adversarial learning can be used to develop high-performing networks trained on unannotated medical images of varying image quality. Specifically, we used low-quality images acquired using inexpensive portable optical systems to train networks for the evaluation of human embryos, the quantification of human sperm morphology and the diagnosis of malarial infections in the blood, and show that the networks performed well across different data distributions. We also show that adversarial learning can be used with unlabelled data from unseen domain-shifted datasets to adapt pretrained supervised networks to new distributions, even when data from the original distribution are not available. Adaptive adversarial networks may expand the use of validated neural-network models for the evaluation of data collected from multiple imaging systems of varying quality without compromising the knowledge stored in the network.
登录
查看更多内容
影响因子：
6.7
作者：
Bormann CL;Thirumalaraju P;Kanakasabapathy MK;Kandula H;Souter I;Dimitriadis I;Gupta R;Pooniwala R;Shafiee H
通讯作者：
Shafiee H
影响因子：
17.1
作者：
Kanakasabapathy MK;Sadasivam M;Singh A;Preston C;Thirumalaraju P;Venkataraman M;Bormann CL;Draz MS;Petrozza JC;Shafiee H
通讯作者：
Shafiee H
影响因子：
7.7
作者：
Bormann CL;Kanakasabapathy MK;Thirumalaraju P;Gupta R;Pooniwala R;Kandula H;Hariton E;Souter I;Dimitriadis I;Ramirez LB;Curchoe CL;Swain J;Boehnlein LM;Shafiee H
通讯作者：
Shafiee H
DOI：
10.1126/science.aay5189
发表时间：
2019-11-22
期刊：
Science (New York, N.Y.)
影响因子：
--
作者：
Hosny A;Aerts HJWL
通讯作者：
Aerts HJWL
影响因子：
9.3
作者：
Kelly, Christopher J.;Karthikesalingam, Alan;King, Dominic
通讯作者：
King, Dominic