Does Medical Imaging learn different Convolution Filters?

TL;DR

This study investigates whether medical imaging models learn distinct convolution filters compared to other domains, finding that differences are mainly due to architecture-specific processing rather than domain-specific learning.

Contribution

The paper demonstrates that medical imaging models do not fundamentally learn different filters, and observed outliers are due to architecture effects rather than domain-specific features.

Findings

Medical imaging models show outlier filter distributions due to architecture.

Standardized architectures trained on medical data are similar to those trained on other data.

Pre-training can be performed with diverse image data without domain-specific constraints.

Abstract

Recent work has investigated the distributions of learned convolution filters through a large-scale study containing hundreds of heterogeneous image models. Surprisingly, on average, the distributions only show minor drifts in comparisons of various studied dimensions including the learned task, image domain, or dataset. However, among the studied image domains, medical imaging models appeared to show significant outliers through "spikey" distributions, and, therefore, learn clusters of highly specific filters different from other domains. Following this observation, we study the collected medical imaging models in more detail. We show that instead of fundamental differences, the outliers are due to specific processing in some architectures. Quite the contrary, for standardized architectures, we find that models trained on medical data do not significantly differ in their filter distributions from similar architectures trained on data from other domains. Our conclusions reinforce previous hypotheses stating that pre-training of imaging models can be done with any kind of diverse image data.