Towards Reducing Segmentation Labeling Costs for CMR Imaging using Explainable AI

Alessa Stria, Asan Agibetov

2022

Abstract

Provided with a sufficient amount of annotated data, deep learning models have been successfully applied to automatically segment cardiac multi-structures from MR images. However, manual delineation of cardiac anatomical structures is expensive to acquire and requires expert knowledge. Recently, weakly- and self-supervised feature learning techniques have been pro-posed to avoid or substantially reduce the effort of manual annotation. Due to their end-to-end design, many of these techniques are hard to train. In this paper, we propose a simple modular segmentation framework based on U-net architecture that injects class activation maps of separately trained classification models to guide the segmentation process. In a small data setting (20-35% of training data), our framework significantly improved the segmentation accuracy of a baseline U-net model (5%-150%).

Paper Citation

in Harvard Style

Stria A. and Agibetov A. (2022). Towards Reducing Segmentation Labeling Costs for CMR Imaging using Explainable AI. In Proceedings of the 1st Workshop on Scarce Data in Artificial Intelligence for Healthcare - Volume 1: SDAIH, ISBN 978-989-758-629-3, SciTePress, pages 11-16. DOI: 10.5220/0011531200003523

in Bibtex Style

@conference{sdaih22,
author={Alessa Stria and Asan Agibetov},
title={Towards Reducing Segmentation Labeling Costs for CMR Imaging using Explainable AI},
booktitle={Proceedings of the 1st Workshop on Scarce Data in Artificial Intelligence for Healthcare - Volume 1: SDAIH,},
year={2022},
pages={11-16},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0011531200003523},
isbn={978-989-758-629-3},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st Workshop on Scarce Data in Artificial Intelligence for Healthcare - Volume 1: SDAIH,
TI - Towards Reducing Segmentation Labeling Costs for CMR Imaging using Explainable AI
SN - 978-989-758-629-3
AU - Stria A.
AU - Agibetov A.
PY - 2022
SP - 11
EP - 16
DO - 10.5220/0011531200003523
PB - SciTePress