Accuracy improvement for Fully Convolutional Networks via selective augmentation with applications to electrocardiogram data

Deep Learning methods have shown suitability for time series classification in the health and medical domain, with promising results for electrocardiogram data classification. Successful identification of myocardial infarction holds lifesaving potential, and any meaningful improvement upon deep learning models in this area is of interest. Conventionally, data augmentation methods are applied universally to the training set when data are limited to ameliorate data resolution or sample size. In the method proposed in this study, data augmentation was not applied in the context of data scarcity. Instead, samples that yielded low confidence predictions from an intermediary test set were selectively augmented to bolster the model's sensitivity to features or patterns less strongly associated with a given class. This approach was tested for improving the performance of a Fully Convolutional Network. The proposed approach achieved 90% accuracy for classifying myocardial infarction as opposed to 82% accuracy for the baseline, a marked improvement. Further, the accuracy of the proposed approach was optimal near a defined upper threshold for qualifying low confidence samples and decreased as this threshold was raised to include higher confidence samples. This suggests exclusively selecting lower confidence samples for data augmentation comes with distinct benefits for electrocardiogram data classification with Fully Convolutional Networks.