Medical Image Quality Assessment based on Probability of Necessity and Sufficiency

TL;DR

This paper introduces a novel medical image quality assessment framework based on the causal inference concept of Probability of Necessity and Sufficiency, improving robustness to out-of-distribution data.

Contribution

It applies PNS to learn features that are both necessary and sufficient for quality prediction, enhancing model robustness in medical image analysis.

Findings

Effective in AS-OCT dataset

Improves robustness to OOD scenarios

Outperforms baseline methods

Abstract

Medical image quality assessment (MIQA) is essential for reliable medical image analysis. While deep learning has shown promise in this field, current models could be misled by spurious correlations learned from data and struggle with out-of-distribution (OOD) scenarios. To that end, we propose an MIQA framework based on a concept from causal inference: Probability of Necessity and Sufficiency (PNS). PNS measures how likely a set of features is to be both necessary (always present for an outcome) and sufficient (capable of guaranteeing an outcome) for a particular result. Our approach leverages this concept by learning hidden features from medical images with high PNS values for quality prediction. This encourages models to capture more essential predictive information, enhancing their robustness to OOD scenarios. We evaluate our framework on an Anterior Segment Optical Coherence Tomography (AS-OCT) dataset for the MIQA task and experimental results demonstrate the effectiveness of our framework.