The Intrinsic Manifolds of Radiological Images and their Role in Deep Learning

TL;DR

This paper investigates the intrinsic manifold structures of radiological images compared to natural images, revealing their impact on deep learning generalization and emphasizing the need for domain-specific models.

Contribution

It provides the first comparative analysis of intrinsic manifold dimensionality in radiological versus natural images and links these properties to generalization challenges in medical imaging.

Findings

Radiological images have lower intrinsic dimensions than natural images.

Generalization ability correlates more strongly with intrinsic dimensionality in medical images.

Radiological images present unique learning challenges requiring tailored deep learning architectures.

Abstract

The manifold hypothesis is a core mechanism behind the success of deep learning, so understanding the intrinsic manifold structure of image data is central to studying how neural networks learn from the data. Intrinsic dataset manifolds and their relationship to learning difficulty have recently begun to be studied for the common domain of natural images, but little such research has been attempted for radiological images. We address this here. First, we compare the intrinsic manifold dimensionality of radiological and natural images. We also investigate the relationship between intrinsic dimensionality and generalization ability over a wide range of datasets. Our analysis shows that natural image datasets generally have a higher number of intrinsic dimensions than radiological images. However, the relationship between generalization ability and intrinsic dimensionality is much stronger for medical images, which could be explained as radiological images having intrinsic features that are more difficult to learn. These results give a more principled underpinning for the intuition that radiological images can be more challenging to apply deep learning to than natural image datasets common to machine learning research. We believe rather than directly applying models developed for natural images to the radiological imaging domain, more care should be taken to developing architectures and algorithms that are more tailored to the specific characteristics of this domain. The research shown in our paper, demonstrating these characteristics and the differences from natural images, is an important first step in this direction.