Beyond Pixels: Medical Image Quality Assessment with Implicit Neural Representations

TL;DR

This paper introduces a novel approach for medical image quality assessment using implicit neural representations (INRs), which handle image variations efficiently and reduce memory usage, improving artifact detection in medical imaging.

Contribution

The work pioneers the use of INRs combined with deep weight space networks, graph neural networks, and transformers for scalable, memory-efficient medical image quality assessment.

Findings

Effective artifact detection on ACDC dataset

Comparable performance with fewer parameters

Robustness to resolution and size variations

Abstract

Artifacts pose a significant challenge in medical imaging, impacting diagnostic accuracy and downstream analysis. While image-based approaches for detecting artifacts can be effective, they often rely on preprocessing methods that can lead to information loss and high-memory-demand medical images, thereby limiting the scalability of classification models. In this work, we propose the use of implicit neural representations (INRs) for image quality assessment. INRs provide a compact and continuous representation of medical images, naturally handling variations in resolution and image size while reducing memory overhead. We develop deep weight space networks, graph neural networks, and relational attention transformers that operate on INRs to achieve image quality assessment. Our method is evaluated on the ACDC dataset with synthetically generated artifact patterns, demonstrating its effectiveness in assessing image quality while achieving similar performance with fewer parameters.