Predicting Scores of Medical Imaging Segmentation Methods with Meta-Learning

TL;DR

This paper explores using meta-learning to predict the performance of medical image segmentation models across various datasets, aiming to streamline model selection with limited resources.

Contribution

It introduces four dataset representations and employs regression models to accurately forecast segmentation performance, demonstrating the potential of meta-learning in medical imaging.

Findings

Meta-learning predicts segmentation performance within 0.10 Dice score on external datasets.

Four dataset representations effectively capture features for performance prediction.

Meta-learning shows promise in aiding model selection in resource-constrained settings.

Abstract

Deep learning has led to state-of-the-art results for many medical imaging tasks, such as segmentation of different anatomical structures. With the increased numbers of deep learning publications and openly available code, the approach to choosing a model for a new task becomes more complicated, while time and (computational) resources are limited. A possible solution to choosing a model efficiently is meta-learning, a learning method in which prior performance of a model is used to predict the performance for new tasks. We investigate meta-learning for segmentation across ten datasets of different organs and modalities. We propose four ways to represent each dataset by meta-features: one based on statistical features of the images and three are based on deep learning features. We use support vector regression and deep neural networks to learn the relationship between the meta-features and prior model performance. On three external test datasets these methods give Dice scores within 0.10 of the true performance. These results demonstrate the potential of meta-learning in medical imaging.