Strategies for Robust Deep Learning Based Deformable Registration

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Honkamaa, Joel, Marttinen, Pekka
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909870848999424
author Honkamaa, Joel
Marttinen, Pekka
author_facet Honkamaa, Joel
Marttinen, Pekka
contents Deep learning based deformable registration methods have become popular in recent years. However, their ability to generalize beyond training data distribution can be poor, significantly hindering their usability. LUMIR brain registration challenge for Learn2Reg 2025 aims to advance the field by evaluating the performance of the registration on contrasts and modalities different from those included in the training set. Here we describe our submission to the challenge, which proposes a very simple idea for significantly improving robustness by transforming the images into MIND feature space before feeding them into the model. In addition, a special ensembling strategy is proposed that shows a small but consistent improvement.
format Preprint
id arxiv_https___arxiv_org_abs_2510_23079
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Strategies for Robust Deep Learning Based Deformable Registration
Honkamaa, Joel
Marttinen, Pekka
Computer Vision and Pattern Recognition
Deep learning based deformable registration methods have become popular in recent years. However, their ability to generalize beyond training data distribution can be poor, significantly hindering their usability. LUMIR brain registration challenge for Learn2Reg 2025 aims to advance the field by evaluating the performance of the registration on contrasts and modalities different from those included in the training set. Here we describe our submission to the challenge, which proposes a very simple idea for significantly improving robustness by transforming the images into MIND feature space before feeding them into the model. In addition, a special ensembling strategy is proposed that shows a small but consistent improvement.
title Strategies for Robust Deep Learning Based Deformable Registration
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2510.23079