重新思考医学图像分类中的迁移学习
Rethinking Transfer Learning for Medical Image Classification
AI总结：
本文提出TruncatedTL，一种差分迁移学习策略，通过重用和微调预训练模型的底层并丢弃其余层，在医学图像分类中实现更优性能与紧凑模型。
摘要：
从预训练深度模型进行迁移学习（TL）是现代医学图像分类（MIC）中的标准做法。然而，要重用哪些级别的特征取决于具体问题，对所有层进行统一的微调可能不是最优的。这一见解部分推动了最近的差分迁移学习策略，如TransFusion（TF）和逐层微调（LWFT），这些策略对预训练模型中的层进行差异化处理。在本文中，我们为该策略家族增加了一种新方法，称为TruncatedTL，它重用并微调适当的底层，并直接丢弃其余层。与其他差分迁移学习方法相比，这不仅带来了更优越的MIC性能，还产生了用于高效推理的紧凑模型。我们的代码可在以下网址获取：https://github.com/sun-umn/TTL
英文摘要：
Transfer learning (TL) from pretrained deep models is a standard practice in modern medical image classification (MIC). However, what levels of features to be reused are problem-dependent, and uniformly finetuning all layers of pretrained models may be suboptimal. This insight has partly motivated the recent differential TL strategies, such as TransFusion (TF) and layer-wise finetuning (LWFT), which treat the layers in the pretrained models differentially. In this paper, we add one more strategy into this family, called TruncatedTL, which reuses and finetunes appropriate bottom layers and directly discards the remaining layers. This yields not only superior MIC performance but also compact models for efficient inference, compared to other differential TL methods. Our code is available at: https://github.com/sun-umn/TTL
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。
