Unsupervised Task Design to Meta-Train Medical Image Classifiers

Meta-training has been empirically demonstrated to be the most effective pre-training method for few-shot learning of medical image classifiers (i.e., classifiers modeled with small training sets). However, the effectiveness of meta-training relies on the availability of a reasonable number of hand-designed classification tasks, which are costly to obtain, and consequently rarely available. In this paper, we propose a new method to unsupervisedly design a large number of classification tasks to meta-train medical image classifiers. We evaluate our method on a breast dynamically contrast enhanced magnetic resonance imaging (DCE-MRI) data set that has been used to benchmark few-shot training methods of medical image classifiers. Our results show that the proposed unsupervised task design to meta-train medical image classifiers builds a pre-trained model that, after fine-tuning, produces better classification results than other unsupervised and supervised pre-training methods, and competitive results with respect to meta-training that relies on hand-designed classification tasks.

[1]  Gustavo Carneiro,et al.  Training Medical Image Analysis Systems like Radiologists , 2018, MICCAI.

[2]  Kilian Q. Weinberger,et al.  Densely Connected Convolutional Networks , 2016, 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR).

[3]  Sergey Levine,et al.  Unsupervised Learning via Meta-Learning , 2018, ICLR.

[4]  Yu-Bin Yang,et al.  Learning Deep Representations Using Convolutional Auto-Encoders with Symmetric Skip Connections , 2016, 2018 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP).

[5]  F J Gilbert,et al.  Personalised screening: is this the way forward? , 2017, Clinical radiology.

[6]  Bram van Ginneken,et al.  A survey on deep learning in medical image analysis , 2017, Medical Image Anal..

[7]  P. Rousseeuw.  Silhouettes: a graphical aid to the interpretation and validation of cluster analysis , 1987 .

[8]  Radford M. Neal.  Pattern Recognition and Machine Learning , 2007, Technometrics.

[9]  N Karssemeijer,et al.  The frequency of missed breast cancers in women participating in a high-risk MRI screening program , 2018, Breast Cancer Research and Treatment.

[10]  Sora C. Yoon,et al.  Interobserver Variability Between Breast Imagers Using the Fifth Edition of the BI-RADS MRI Lexicon. , 2015, AJR. American journal of roentgenology.

[11]  Nasser M. Nasrabadi,et al.  Pattern Recognition and Machine Learning , 2006, Technometrics.

[12]  Lina Arbash Meinel,et al.  Breast MRI lesion classification: Improved performance of human readers with a backpropagation neural network computer‐aided diagnosis (CAD) system , 2007, Journal of magnetic resonance imaging : JMRI.

[13]  Stuart Crozier,et al.  Fully automatic lesion segmentation in breast MRI using mean‐shift and graph‐cuts on a region adjacency graph , 2014, Journal of magnetic resonance imaging : JMRI.

[14]  Matthijs Douze,et al.  Deep Clustering for Unsupervised Learning of Visual Features , 2018, ECCV.

[15]  Gustavo Carneiro,et al.  Pre and Post-hoc Diagnosis and Interpretation of Malignancy from Breast DCE-MRI , 2018, Medical Image Anal..

[16]  Sergey Levine,et al.  Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks , 2017, ICML.

[17]  Andrew P. Bradley,et al.  The use of the area under the ROC curve in the evaluation of machine learning algorithms , 1997, Pattern Recognit..

[18]  Xiaohui Xie,et al.  Deep Multi-instance Networks with Sparse Label Assignment for Whole Mammogram Classification , 2016, bioRxiv.

[19]  Wufeng Xue,et al.  Full left ventricle quantification via deep multitask relationships learning , 2018, Medical Image Anal..

[20]  Thomas L. Griffiths,et al.  Recasting Gradient-Based Meta-Learning as Hierarchical Bayes , 2018, ICLR.

[21]  Hayit Greenspan,et al.  Chest pathology detection using deep learning with non-medical training , 2015, 2015 IEEE 12th International Symposium on Biomedical Imaging (ISBI).