Improved Trainable Calibration Method for Neural Networks on Medical Imaging Classification

TL;DR

This paper introduces a new calibration method for neural networks in medical imaging that improves uncertainty estimates without sacrificing classification accuracy, enhancing safety in medical decision systems.

Contribution

A novel calibration technique based on expected calibration error that can be integrated into training, improving model calibration across architectures and datasets.

Findings

Significantly reduces calibration error in neural networks.

Maintains high classification accuracy.

Easily integrated as an auxiliary loss during training.

Abstract

Recent works have shown that deep neural networks can achieve super-human performance in a wide range of image classification tasks in the medical imaging domain. However, these works have primarily focused on classification accuracy, ignoring the important role of uncertainty quantification. Empirically, neural networks are often miscalibrated and overconfident in their predictions. This miscalibration could be problematic in any automatic decision-making system, but we focus on the medical field in which neural network miscalibration has the potential to lead to significant treatment errors. We propose a novel calibration approach that maintains the overall classification accuracy while significantly improving model calibration. The proposed approach is based on expected calibration error, which is a common metric for quantifying miscalibration. Our approach can be easily integrated into any classification task as an auxiliary loss term, thus not requiring an explicit training round for calibration. We show that our approach reduces calibration error significantly across various architectures and datasets.