Flow-Mixup: Classifying Multi-labeled Medical Images with Corrupted Labels

Abstract

In clinical practice, medical image interpretation often involves multi-labeled classification, since the affected parts of a patient tend to present multiple symptoms or comorbidities. Recently, deep learning based frameworks have attained expertlevel performance on medical image interpretation, which can be attributed partially to large amounts of accurate annotations. However, manually annotating massive amounts of medical images is impractical, while automatic annotation is fast but imprecise (possibly introducing corrupted labels). In this work, we propose a new regularization approach, called Flow-Mixup, for multi-labeled medical image classification with corrupted labels. Flow-Mixup guides the models to capture robust features for each abnormality, thus helping handle corrupted labels effectively and making it possible to apply automatic annotation. Specifically, Flow-Mixup decouples the extracted features by adding constraints to the hidden states of the models. Also, Flow-Mixup is more stable and effective comparing to other known regularization methods, as shown by theoretical and empirical analyses. Experiments on two electrocardiogram datasets and a chest X-ray dataset containing corrupted labels verify that Flow-Mixup is effective and insensitive to corrupted labels.

Identifying Electrocardiogram Abnormalities Using a Handcrafted-Rule-Enhanced Neural Network

“…As previous research has demonstrated that CNNs can well process ECG signals [13], [21], we develop an end-to-end CNN to cope with ECG signals for arrhythmia classification.…”

“…In early years, various traditional methods employing decision trees [15], SVM [16], random forests [17], and Bayesian networks [18] were applied to classify ECG signals, but did not yield satisfactory performances. Recently, deep learning approaches have drastically improved performances of various recognition tasks, including automatic ECG diagnosis [10], [13], [19]- [22]. Deep learning methods for ECG can be roughly divided into three types, graph based [23]- [25], recurrent neural network (RNN) based [26]- [28], and convolution based [13], [14], [21], [29] methods.…”

“…Recently, deep learning approaches have drastically improved performances of various recognition tasks, including automatic ECG diagnosis [10], [13], [19]- [22]. Deep learning methods for ECG can be roughly divided into three types, graph based [23]- [25], recurrent neural network (RNN) based [26]- [28], and convolution based [13], [14], [21], [29] methods. Specifically, graph based neural networks aim to capture the dependencies among cardiac abnormalities, since many ECG cases belong to multiple categories (abnormality types).…”

Electrocardio Panorama: Synthesizing New ECG views with Self-supervision

“…. , L}) with L views, we separately deal with these ECG views in parallel by group convolutions, instead of fusing them together as in the previous work [Kachuee et al, 2018;Chen et al, 2020b]. Formally, we predict L basic representations and L groups of deflection representations, and then compute the averaged basic representations Z b and averaged deflection representations Z d over the L views as the input of the decoder.…”

“…Motivated by representation interpolations [Karras et al, 2019] and Mixup [Zhang et al, 2018;Chen et al, 2020b;Chen et al, 2020a], we synthesize electrocardio field representations of a new ECG by mixing the extracted electrocardio field representations of the same categories (e.g., diseases). As the range of each deflection type varies and the features of different types of deflections should not be mixed for new data synthesis, for this task we train a Nef-Net such that the electrocardio field representation is only predicted by the deflection representation learning branch (i.e., removing the branch for the basic representation).…”