Leveraging Self-supervised Learning for Fetal Cardiac Planes Classification Using Ultrasound Scan Videos

Authors

Joseph Geo Benjamin, Mohamed Bin Zayed University of Artificial IntelligenceFollow
Mothilal Asokan, Mohamed Bin Zayed University of Artificial IntelligenceFollow
Amna Alhosani, Mohamed Bin Zayed University of Artificial IntelligenceFollow
Hussain Alasmawi, Mohamed Bin Zayed University of Artificial IntelligenceFollow
Werner Gerhard Diehl, Abu Dhabi Health Services Company
Leanne Bricker, Abu Dhabi Health Services Company
Karthik Nandakumar, Mohamed Bin Zayed University of Artificial IntelligenceFollow
Mohammad Yaqub, Mohamed Bin Zayed University of Artificial IntelligenceFollow

Document Type

Conference Proceeding

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Abstract

Self-supervised learning (SSL) methods are popular since they can address situations with limited annotated data by directly utilising the underlying data distribution. However, adoption of such methods is not explored enough in ultrasound (US) imaging, especially for fetal assessment. We investigate the potential of dual-encoder SSL in utilizing unlabelled US video data to improve the performance of challenging downstream Standard Fetal Cardiac Planes (SFCP) classification using limited labelled 2D US images. We study 7 SSL approaches based on reconstruction, contrastive loss, distillation and information theory, and evaluate them extensively on a large private US dataset. Our observations and finding are consolidated from more than 500 downstream training experiments under different settings. Our primary observation shows that for SSL training, the variance of the dataset is more crucial than its size because it allows the model to learn generalisable representations which improve the performance of downstream tasks. Overall, the BarlowTwins method shows robust performance irrespective of the training settings and data variations, when used as an initialisation for downstream tasks. Notably, full fine-tuning with 1 % of labelled data outperforms ImageNet initialisation by 12 % in F1-score and outperforms other SSL initialisations by at least 4 % in F1-score, thus making it a promising candidate for transfer learning from US video to image data. Our code is available at https://github.com/BioMedIA-MBZUAI/Ultrasound-SSL-FetalCardiacPlanes.

First Page

68

Last Page

78

DOI

10.1007/978-3-031-44521-7_7

Publication Date

10-1-2023

Keywords

Self-supervised Learning, Standard Fetal Cardiac Planes, Ultrasound Scan Videos

Comments

IR conditions: non-described

Recommended Citation

J.G. Benjamin et al., "Leveraging Self-supervised Learning for Fetal Cardiac Planes Classification Using Ultrasound Scan Videos". In 4th Intl. Workshop of Advances in Simplifying Medical Ultrasound (ASMUS 2023) LNCS, Springer, Cham., vol 14337, pp. 68-78, 2023. doi:10.1007/978-3-031-44521-7_7

Additional Links

https://doi.org/10.1007/978-3-031-44521-7_7