Rethink Transfer Learning in Medical Image Classification [preprint]
Preprint date
June 9, 2021
Authors
Le Peng (Ph.D. student), Hengyue Liang, Taihui Li (Ph.D. student), Ju Sun (assistant professor)
Abstract
Transfer learning (TL) with deep convolutional neural networks (DCNNs) has proved successful in medical image classification (MIC). However, the current practice is puzzling, as MIC typically relies only on low- and/or mid-level features that are learned in the bottom layers of DCNNs. Following this intuition, we question the current strategies of TL in MIC. In this paper, we perform careful experimental comparisons between shallow and deep networks for classification on two chest x-ray datasets, using different TL strategies. We find that deep models are not always favorable, and finetuning truncated deep models almost always yields the best performance, especially in data-poor regimes.
Link to full paper
Rethink Transfer Learning in Medical Image Classification
Keywords
transfer learning, machine learning, medical imaging