DOMINO: Domain-aware Model Calibration in Medical Image Segmentation
This addresses the problem of unreliable model predictions in high-risk medical image segmentation for clinicians, though it appears incremental as it builds on existing calibration techniques with domain-specific adaptations.
The paper tackled poor calibration in deep neural networks for medical image segmentation, which undermines trustworthiness, by introducing DOMINO, a domain-aware calibration method that leverages semantic confusability and hierarchical similarity between class labels; results showed it consistently achieved better calibration, higher accuracy, and faster inference times than non-calibrated and state-of-the-art methods, especially on rarer classes.
Model calibration measures the agreement between the predicted probability estimates and the true correctness likelihood. Proper model calibration is vital for high-risk applications. Unfortunately, modern deep neural networks are poorly calibrated, compromising trustworthiness and reliability. Medical image segmentation particularly suffers from this due to the natural uncertainty of tissue boundaries. This is exasperated by their loss functions, which favor overconfidence in the majority classes. We address these challenges with DOMINO, a domain-aware model calibration method that leverages the semantic confusability and hierarchical similarity between class labels. Our experiments demonstrate that our DOMINO-calibrated deep neural networks outperform non-calibrated models and state-of-the-art morphometric methods in head image segmentation. Our results show that our method can consistently achieve better calibration, higher accuracy, and faster inference times than these methods, especially on rarer classes. This performance is attributed to our domain-aware regularization to inform semantic model calibration. These findings show the importance of semantic ties between class labels in building confidence in deep learning models. The framework has the potential to improve the trustworthiness and reliability of generic medical image segmentation models. The code for this article is available at: https://github.com/lab-smile/DOMINO.