| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 3.33 MB | Adobe PDF |
Authors
Abstract(s)
Self-supervised learning has proven to be an effective way to learn representations in domains
where annotated labels are scarce, such as medical imaging. A widely adopted framework for this
purpose is contrastive learning and it has been applied to different scenarios. This thesis seeks
to advance our understanding of the contrastive learning framework by exploring a novel perspective: employing multi-organ datasets for pre-training models tailored to specific organ-related
target tasks. More specifically, our target task is breast tumour segmentation in ultrasound images.
The pre-training datasets include ultrasound images from other organs, such as the lungs and heart,
and large datasets of natural images. Our results show that conventional contrastive learning pretraining improves performance compared to supervised baseline approaches. Furthermore, our
pre-trained models achieve comparable performance when fine-tuned with only half of the available labelled data. Our findings also show the advantages of pre-training on diverse organ data for
improving performance in the downstream task.
Description
Tese de Mestrado, Engenharia Informática, 2024, Universidade de Lisboa, Faculdade de Ciências
Keywords
Aprendizagem Auto-Supervisionada Aprendizagem Contrastiva Segmentação de Imagens Imagens de Ultrassons Segmentação de tumor de mama Teses de mestrado - 2024