Automatic segmentation with detection of local segmentation failures in cardiac MRI

TL;DR

This paper presents a method combining CNN-based cardiac MRI segmentation with uncertainty estimation to detect local segmentation failures, enabling targeted manual correction and improved overall accuracy.

Contribution

It introduces a novel approach that integrates segmentation uncertainty measures with a failure detection CNN to enhance cardiac MRI segmentation robustness.

Findings

Combining automatic segmentation with failure detection improves accuracy.

Uncertainty measures like entropy and MC-dropout effectively identify segmentation failures.

Simulated manual correction of detected failures significantly boosts segmentation performance.

Abstract

Segmentation of cardiac anatomical structures in cardiac magnetic resonance images (CMRI) is a prerequisite for automatic diagnosis and prognosis of cardiovascular diseases. To increase robustness and performance of segmentation methods this study combines automatic segmentation and assessment of segmentation uncertainty in CMRI to detect image regions containing local segmentation failures. Three state-of-the-art convolutional neural networks (CNN) were trained to automatically segment cardiac anatomical structures and obtain two measures of predictive uncertainty: entropy and a measure derived by MC-dropout. Thereafter, using the uncertainties another CNN was trained to detect local segmentation failures that potentially need correction by an expert. Finally, manual correction of the detected regions was simulated. Using publicly available CMR scans from the MICCAI 2017 ACDC challenge, the impact of CNN architecture and loss function for segmentation, and the uncertainty measure was investigated. Performance was evaluated using the Dice coefficient and 3D Hausdorff distance between manual and automatic segmentation. The experiments reveal that combining automatic segmentation with simulated manual correction of detected segmentation failures leads to statistically significant performance increase.