Intra-video Positive Pairs in Self-Supervised Learning for Ultrasound
This work addresses data scarcity in medical imaging, specifically for ultrasound analysis, by introducing a domain-specific enhancement to self-supervised learning, though it is incremental as it builds on existing contrastive methods.
The study tackled the problem of limited labeled data in medical imaging by proposing Intra-Video Positive Pairs (IVPP), a self-supervised learning method that uses proximal images from the same ultrasound video as pairs, resulting in an average test accuracy improvement of ≥1.3% on COVID-19 classification with the POCUS dataset.
Self-supervised learning (SSL) is one strategy for addressing the paucity of labelled data in medical imaging by learning representations from unlabelled images. Contrastive and non-contrastive SSL methods produce learned representations that are similar for pairs of related images. Such pairs are commonly constructed by randomly distorting the same image twice. The videographic nature of ultrasound offers flexibility for defining the similarity relationship between pairs of images. In this study, we investigated the effect of utilizing proximal, distinct images from the same B-mode ultrasound video as pairs for SSL. Additionally, we introduced a sample weighting scheme that increases the weight of closer image pairs and demonstrated how it can be integrated into SSL objectives. Named Intra-Video Positive Pairs (IVPP), the method surpassed previous ultrasound-specific contrastive learning methods' average test accuracy on COVID-19 classification with the POCUS dataset by $\ge 1.3\%$. Detailed investigations of IVPP's hyperparameters revealed that some combinations of IVPP hyperparameters can lead to improved or worsened performance, depending on the downstream task. Guidelines for practitioners were synthesized based on the results, such as the merit of IVPP with task-specific hyperparameters, and the improved performance of contrastive methods for ultrasound compared to non-contrastive counterparts.