Incremental Learning Meets Transfer Learning: Application to Multi-site Prostate MRI Segmentation.
Incremental Learning Meets Transfer Learning: Application to Multi-site Prostate MRI Segmentation.
复制标题
渐进学习与迁移学习的结合：在多部位前列腺 MRI 分割中的应用。
DOI：
10.1007/978-3-031-18523-6_1
复制
发表时间：
2022
期刊：
影响因子：
--
通讯作者：
Duncan,JamesS
中科院分区：
文献类型：
--
作者：
You,Chenyu;Xiang,Jinlin;Su,Kun;Zhang,Xiaoran;Dong,Siyuan;Onofrey,John;Staib,Lawrence;Duncan,JamesS
最近已经创建了许多医学数据集用于医学图像分割任务，并且很自然地质疑我们是否可以使用它们依次训练单个模型(1)在所有这些数据集上表现更好，(2)泛化良好并更好地转移到未知的目标位点域。先前的工作通过在多站点数据集上联合训练一个模型来实现这一目标，平均而言达到了竞争性能，但这种方法依赖于对所有训练数据可用性的假设，从而限制了其在实际部署中的有效性。在本文中，我们提出了一种新的多站点分割框架，称为增量迁移学习（ITL），它以端到端顺序的方式从多站点数据集中学习模型。具体来说，“增量”指的是训练顺序构建的数据集，而“转移”是通过利用每个数据集上嵌入特征的线性组合中的有用信息来实现的。此外，我们介绍了我们的ITL框架，我们在其中训练网络，包括一个具有预训练权重的站点不可知编码器和最多两个分割解码器头。为了更好地泛化目标域，我们还设计了一种新的站点级增量损失。其次，我们首次证明了利用我们的ITL训练方案能够缓解增量学习中具有挑战性的灾难性遗忘问题。我们使用五个具有挑战性的基准数据集进行实验，以验证我们的增量迁移学习方法的有效性。我们的方法对计算资源和特定领域的专业知识的假设最小，因此构成了多站点医学图像分割的一个强有力的起点。
Many medical datasets have recently been created for medical image segmentation tasks, and it is natural to question whether we can use them to sequentially train a single model that (1) performs better on all these datasets, and (2) generalizes well and transfers better to the unknown target site domain. Prior works have achieved this goal by jointly training one model on multi-site datasets, which achieve competitive performance on average but such methods rely on the assumption about the availability of all training data, thus limiting its effectiveness in practical deployment. In this paper, we propose a novel multi-site segmentation framework calledincremental-transfer learning (ITL), which learns a model from multi-site datasets in an end-to-end sequential fashion. Specifically, “incremental” refers to training sequentially constructed datasets, and “transfer” is achieved by leveraging useful information from the linear combination of embedding features on each dataset. In addition, we introduce our ITL framework, where we train the network including a site-agnostic encoder with pretrained weights and at most two segmentation decoder heads. We also design a novel site-level incremental loss in order to generalize well on the target domain. Second, we show for the first time that leveraging our ITL training scheme is able to alleviate challenging catastrophic forgetting problems in incremental learning. We conduct experiments using five challenging benchmark datasets to validate the effectiveness of our incremental-transfer learning approach. Our approach makes minimal assumptions on computation resources and domain-specific expertise, and hence constitutes a strong starting point in multi-site medical image segmentation.
登录
查看更多内容
影响因子：
--
作者：
W. Miall;P. D. Oldham
通讯作者：
P. D. Oldham
A 型行为模式的家庭方面和对压力的生理反应
影响因子：
168.9
作者：
K. Matthews;C. J. Rakaczky
通讯作者：
C. J. Rakaczky
冷加压试验和基础血压的预后价值。
影响因子：
2.8
作者：
W. Harlan;R. K. Osborne;Ashton Graybiel
通讯作者：
Ashton Graybiel
影响因子：
4.2
作者：
Glass David C.;Lake C. Raymond;Contrada Richard J.;Kehoe Kathleen;Erlanger Laura R.
通讯作者：
Erlanger Laura R.
双胞胎心血管应激反应的相似性。
DOI：
10.1037//0278-6133.4.5.413
发表时间：
1985
期刊：
Health psychology : official journal of the Division of Health Psychology, American Psychological Association
影响因子：
--
作者：
Carmelli,D;Chesney,MA;Ward,MM;Rosenman,RH
通讯作者：
Rosenman,RH