End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

Jiandong Wang; Alessandro Perelli

doi:10.48550/arXiv.2406.00479

End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

Jiandong Wang
, Alessandro Perelli

Biomedical Engineering

Research output: Working paper/Preprint › Preprint

129 Downloads (Pure)

Abstract

Dual energy X-ray Computed Tomography (DECT) enables to automatically decompose materials in clinical images without the manual segmentation using the dependency of the X-ray linear attenuation with energy. In this work we propose a deep learning procedure called End-to-End Material Decomposition (E2E-DEcomp) for quantitative material decomposition which directly convert the CT projection data into material images. The algorithm is based on incorporating the knowledge of the spectral model DECT system into the deep learning training loss and combining a data-learned prior in the material image domain. Furthermore, the training does not require any energy-based images in the dataset but rather only sinogram and material images. We show the effectiveness of the proposed direct E2E-DEcomp method on the AAPM spectral CT dataset (Sidky and Pan, 2023) compared with state of the art supervised deep learning networks.

Original language	English
Publisher	arXiv
Number of pages	7
DOIs	https://doi.org/10.48550/arXiv.2406.00479
Publication status	Published - 1 Jun 2024

Access to Document

10.48550/arXiv.2406.00479Licence: CC BY

2406.00479v1Submitted manuscript, 497 KBLicence: CC BY

End-to-End Model-Based Deep Learning for Dual-Energy Computed Tomography Material Decomposition
Wang, J. & Perelli, A., 22 Aug 2024, 2024 IEEE International Symposium on Biomedical Imaging (ISBI). IEEE, 5 p. (Proceedings - International Symposium on Biomedical Imaging).
Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
1 Link opens in a new tab Citation (Scopus)

Cite this

@techreport{ca0f991b9fd14a07ba29119539cd13b7,

title = "End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition",

abstract = "Dual energy X-ray Computed Tomography (DECT) enables to automatically decompose materials in clinical images without the manual segmentation using the dependency of the X-ray linear attenuation with energy. In this work we propose a deep learning procedure called End-to-End Material Decomposition (E2E-DEcomp) for quantitative material decomposition which directly convert the CT projection data into material images. The algorithm is based on incorporating the knowledge of the spectral model DECT system into the deep learning training loss and combining a data-learned prior in the material image domain. Furthermore, the training does not require any energy-based images in the dataset but rather only sinogram and material images. We show the effectiveness of the proposed direct E2E-DEcomp method on the AAPM spectral CT dataset (Sidky and Pan, 2023) compared with state of the art supervised deep learning networks.",

author = "Jiandong Wang and Alessandro Perelli",

year = "2024",

month = jun,

day = "1",

doi = "10.48550/arXiv.2406.00479",

language = "English",

publisher = "arXiv",

type = "WorkingPaper",

institution = "arXiv",

}

TY - UNPB

T1 - End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

AU - Wang, Jiandong

AU - Perelli, Alessandro

PY - 2024/6/1

Y1 - 2024/6/1

N2 - Dual energy X-ray Computed Tomography (DECT) enables to automatically decompose materials in clinical images without the manual segmentation using the dependency of the X-ray linear attenuation with energy. In this work we propose a deep learning procedure called End-to-End Material Decomposition (E2E-DEcomp) for quantitative material decomposition which directly convert the CT projection data into material images. The algorithm is based on incorporating the knowledge of the spectral model DECT system into the deep learning training loss and combining a data-learned prior in the material image domain. Furthermore, the training does not require any energy-based images in the dataset but rather only sinogram and material images. We show the effectiveness of the proposed direct E2E-DEcomp method on the AAPM spectral CT dataset (Sidky and Pan, 2023) compared with state of the art supervised deep learning networks.

AB - Dual energy X-ray Computed Tomography (DECT) enables to automatically decompose materials in clinical images without the manual segmentation using the dependency of the X-ray linear attenuation with energy. In this work we propose a deep learning procedure called End-to-End Material Decomposition (E2E-DEcomp) for quantitative material decomposition which directly convert the CT projection data into material images. The algorithm is based on incorporating the knowledge of the spectral model DECT system into the deep learning training loss and combining a data-learned prior in the material image domain. Furthermore, the training does not require any energy-based images in the dataset but rather only sinogram and material images. We show the effectiveness of the proposed direct E2E-DEcomp method on the AAPM spectral CT dataset (Sidky and Pan, 2023) compared with state of the art supervised deep learning networks.

U2 - 10.48550/arXiv.2406.00479

DO - 10.48550/arXiv.2406.00479

M3 - Preprint

BT - End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

PB - arXiv

ER -

End-to-End Model-based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

Abstract

Access to Document

Fingerprint

Research output

End-to-End Model-Based Deep Learning for Dual-Energy Computed Tomography Material Decomposition

Cite this