Scale-equivariant deep model-based optoacoustic image reconstruction

Model-based reconstruction provides state-of-the-art image quality for multispectral optoacoustic tomography. However, optimal regularization of in vivo data necessitates scan-specific adjustments of the regularization strength to compensate for fluctuations of the signal magnitudes between differen...

Full description

Saved in:
Bibliographic Details
Main Authors: Christoph Dehner, Ledia Lilaj, Vasilis Ntziachristos, Guillaume Zahnd, Dominik Jüstel
Format: Article
Language:English
Published: Elsevier 2025-08-01
Series:Photoacoustics
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S2213597925000503
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849419612165242880
author Christoph Dehner
Ledia Lilaj
Vasilis Ntziachristos
Guillaume Zahnd
Dominik Jüstel
author_facet Christoph Dehner
Ledia Lilaj
Vasilis Ntziachristos
Guillaume Zahnd
Dominik Jüstel
author_sort Christoph Dehner
collection DOAJ
description Model-based reconstruction provides state-of-the-art image quality for multispectral optoacoustic tomography. However, optimal regularization of in vivo data necessitates scan-specific adjustments of the regularization strength to compensate for fluctuations of the signal magnitudes between different sinograms. Magnitude fluctuations within in vivo data also pose a challenge for supervised deep learning of a model-based reconstruction operator, as training data must cover the complete range of expected signal magnitudes. In this work, we derive a scale-equivariant model-based reconstruction operator that i) automatically adjusts the regularization strength based on the L2 norm of the input sinogram, and ii) facilitates supervised deep learning of the operator using input singorams with a fixed norm. Scale-equivariant model-based reconstruction applies appropriate regularization to sinograms of arbitrary magnitude, achieves slightly better accuracy in quantifying blood oxygen saturation, and enables more accurate supervised deep learning of the operator.
format Article
id doaj-art-e1867903d308445fbd5d45b6c4ba3a17
institution Kabale University
issn 2213-5979
language English
publishDate 2025-08-01
publisher Elsevier
record_format Article
series Photoacoustics
spelling doaj-art-e1867903d308445fbd5d45b6c4ba3a172025-08-20T03:32:01ZengElsevierPhotoacoustics2213-59792025-08-014410072710.1016/j.pacs.2025.100727Scale-equivariant deep model-based optoacoustic image reconstructionChristoph Dehner0Ledia Lilaj1Vasilis Ntziachristos2Guillaume Zahnd3Dominik Jüstel4iThera Medical GmbH, Munich, Germany; Institute of Biological and Medical Imaging, Helmholtz Zentrum München, Neuherberg, GermanyiThera Medical GmbH, Munich, GermanyInstitute of Biological and Medical Imaging, Helmholtz Zentrum München, Neuherberg, Germany; Chair of Biological Imaging at the Central Institute for Translational Cancer Research (TranslaTUM), School of Medicine and Health, Technical University of Munich, Munich, Germany; Munich Institute of Robotics and Machine Intelligence (MIRMI), Technical University of Munich, Munich, GermanyiThera Medical GmbH, Munich, Germany; Institute of Biological and Medical Imaging, Helmholtz Zentrum München, Neuherberg, GermanyInstitute of Biological and Medical Imaging, Helmholtz Zentrum München, Neuherberg, Germany; Chair of Biological Imaging at the Central Institute for Translational Cancer Research (TranslaTUM), School of Medicine and Health, Technical University of Munich, Munich, Germany; Institute of Computational Biology, Helmholtz Zentrum München, Neuherberg, Germany; Corresponding author at: Chair of Biological Imaging at the Central Institute for Translational Cancer Research (TranslaTUM), School of Medicine and Health, Technical University of Munich, Munich, Germany.Model-based reconstruction provides state-of-the-art image quality for multispectral optoacoustic tomography. However, optimal regularization of in vivo data necessitates scan-specific adjustments of the regularization strength to compensate for fluctuations of the signal magnitudes between different sinograms. Magnitude fluctuations within in vivo data also pose a challenge for supervised deep learning of a model-based reconstruction operator, as training data must cover the complete range of expected signal magnitudes. In this work, we derive a scale-equivariant model-based reconstruction operator that i) automatically adjusts the regularization strength based on the L2 norm of the input sinogram, and ii) facilitates supervised deep learning of the operator using input singorams with a fixed norm. Scale-equivariant model-based reconstruction applies appropriate regularization to sinograms of arbitrary magnitude, achieves slightly better accuracy in quantifying blood oxygen saturation, and enables more accurate supervised deep learning of the operator.http://www.sciencedirect.com/science/article/pii/S2213597925000503Optoacoustic imagingModel-based reconstructionRegularizationScale-equivariance
spellingShingle Christoph Dehner
Ledia Lilaj
Vasilis Ntziachristos
Guillaume Zahnd
Dominik Jüstel
Scale-equivariant deep model-based optoacoustic image reconstruction
Photoacoustics
Optoacoustic imaging
Model-based reconstruction
Regularization
Scale-equivariance
title Scale-equivariant deep model-based optoacoustic image reconstruction
title_full Scale-equivariant deep model-based optoacoustic image reconstruction
title_fullStr Scale-equivariant deep model-based optoacoustic image reconstruction
title_full_unstemmed Scale-equivariant deep model-based optoacoustic image reconstruction
title_short Scale-equivariant deep model-based optoacoustic image reconstruction
title_sort scale equivariant deep model based optoacoustic image reconstruction
topic Optoacoustic imaging
Model-based reconstruction
Regularization
Scale-equivariance
url http://www.sciencedirect.com/science/article/pii/S2213597925000503
work_keys_str_mv AT christophdehner scaleequivariantdeepmodelbasedoptoacousticimagereconstruction
AT ledialilaj scaleequivariantdeepmodelbasedoptoacousticimagereconstruction
AT vasilisntziachristos scaleequivariantdeepmodelbasedoptoacousticimagereconstruction
AT guillaumezahnd scaleequivariantdeepmodelbasedoptoacousticimagereconstruction
AT dominikjustel scaleequivariantdeepmodelbasedoptoacousticimagereconstruction