Semantic similarity metrics for learned image registration

Steffen Czolbe, Oswin Krause, Aasa Feragen

Keywords: Image Registration, Deep Learning, Feature Embedding

TL;DR: Novel semantic loss function for learned image registration. Dataset-specific semantic features are learned in both unsupervised and supervised settings. Somewhat invariant to noise and artifacts, beating baselines.

Abstract: We propose a semantic similarity metric for image registration. Existing metrics like euclidean distance or normalized cross-correlation focus on aligning intensity values, giving difficulties with low intensity contrast or noise. Our approach learns dataset-specific features that drive the optimization of a learning-based registration model. We train both an unsupervised approach using an auto-encoder, and a semi-supervised approach using supplemental segmentation data to extract semantic features for image registration. Comparing to existing methods across multiple image modalities and applications, we achieve consistently high registration accuracy. A learned invariance to noise gives smoother transformations on low-quality images.

Registration: I acknowledge that publication of this at MIDL and in the proceedings requires at least one of the authors to register and present the work during the conference.

Authorship: I confirm that I am the author of this work and that it has not been submitted to another publication before.

Paper Type: methodological development

Primary Subject Area: Image Registration

Secondary Subject Area: Unsupervised Learning and Representation Learning

Source Code Url: https://github.com/SteffenCzolbe/DeepSimRegistration

Data Set Url: http://celltrackingchallenge.net/2d-datasets/, http://fcon_1000.projects.nitrc.org/indi/abide/, https://www.oasis-brains.org/

Community Implementations: [ 1 code implementation](https://www.catalyzex.com/paper/arxiv:2104.10051/code)