PhD Seminar • Artificial Intelligence • Intra-video Positive Pairs in Self-Supervised Learning for Ultrasound

Please note: This PhD seminar will take place in DC 2310.

Blake VanBerlo, PhD candidate
David R. Cheriton School of Computer Science

Supervisors: Professors Jesse Hoey, Alexander Wong

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 learn 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, the method surpassed previous ultrasound-specific contrastive learning methods’ average test accuracy on COVID-19 classification with the POCUS dataset by ≥ 1.3%. Investigations revealed that some combinations of IVUP 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 IVUP with task-specific hyperparameters, and the improved performance of contrastive methods for ultrasound compared to non-contrastive counterparts.