Improving Medical Image Classification in Noisy Labels Using Only Self-supervised Pretraining
This addresses noisy label issues in medical imaging, where datasets are small and variations subtle, but it is incremental as it extends known methods from natural images to medical domains.
The paper tackled the problem of noisy labels in medical image classification by exploring self-supervised pretraining methods, showing that models initialized with such weights learn better features and improve robustness, with results demonstrated on NCT-CRC-HE-100K and COVID-QU-Ex datasets.
Noisy labels hurt deep learning-based supervised image classification performance as the models may overfit the noise and learn corrupted feature extractors. For natural image classification training with noisy labeled data, model initialization with contrastive self-supervised pretrained weights has shown to reduce feature corruption and improve classification performance. However, no works have explored: i) how other self-supervised approaches, such as pretext task-based pretraining, impact the learning with noisy label, and ii) any self-supervised pretraining methods alone for medical images in noisy label settings. Medical images often feature smaller datasets and subtle inter class variations, requiring human expertise to ensure correct classification. Thus, it is not clear if the methods improving learning with noisy labels in natural image datasets such as CIFAR would also help with medical images. In this work, we explore contrastive and pretext task-based self-supervised pretraining to initialize the weights of a deep learning classification model for two medical datasets with self-induced noisy labels -- NCT-CRC-HE-100K tissue histological images and COVID-QU-Ex chest X-ray images. Our results show that models initialized with pretrained weights obtained from self-supervised learning can effectively learn better features and improve robustness against noisy labels.