A Deep Learning Approach to MR-less Spatial Normalization for Tau PET Images
(2019) 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019 In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 11765 LNCS. p.355-363- Abstract
The procedure of aligning a positron emission tomography (PET) image with a common coordinate system, spatial normalization, typically demands a corresponding structural magnetic resonance (MR) image. However, MR imaging is not always available or feasible for the subject, which calls for enabling spatial normalization without MR, MR-less spatial normalization. In this work, we propose a template-free approach to MR-less spatial normalization for [18F]flortaucipir tau PET images. We use a deep neural network that estimates an aligning transformation from the PET input image, and outputs the spatially normalized image as well as the parameterized transformation. In order to do so, the proposed network iteratively estimates a set of rigid... (More)
The procedure of aligning a positron emission tomography (PET) image with a common coordinate system, spatial normalization, typically demands a corresponding structural magnetic resonance (MR) image. However, MR imaging is not always available or feasible for the subject, which calls for enabling spatial normalization without MR, MR-less spatial normalization. In this work, we propose a template-free approach to MR-less spatial normalization for [18F]flortaucipir tau PET images. We use a deep neural network that estimates an aligning transformation from the PET input image, and outputs the spatially normalized image as well as the parameterized transformation. In order to do so, the proposed network iteratively estimates a set of rigid and affine transformations by means of convolutional neural network regressors as well as spatial transformer layers. The network is trained and validated on 199 tau PET volumes with corresponding ground truth transformations, and tested on two different datasets. The proposed method shows competitive performance in terms of registration accuracy as well as speed, and compares favourably to previously published results.
(Less)
- author
- Alvén, Jennifer ; Heurling, Kerstin ; Smith, Ruben LU ; Strandberg, Olof LU ; Schöll, Michael LU ; Hansson, Oskar LU and Kahl, Fredrik
- organization
- publishing date
- 2019-10-10
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Medical Image Computing and Computer Assisted Intervention : MICCAI 2019 - 22nd International Conference, Proceedings - MICCAI 2019 - 22nd International Conference, Proceedings
- series title
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- editor
- Shen, Dinggang ; Yap, Pew-Thian ; Liu, Tianming ; Peters, Terry M. ; Khan, Ali ; Staib, Lawrence H. ; Essert, Caroline and Zhou, Sean
- volume
- 11765 LNCS
- pages
- 9 pages
- publisher
- Springer Nature
- conference name
- 22nd International Conference on Medical Image Computing and Computer-Assisted Intervention, MICCAI 2019
- conference location
- Shenzhen, China
- conference dates
- 2019-10-13 - 2019-10-17
- external identifiers
-
- scopus:85075683154
- ISSN
- 1611-3349
- 0302-9743
- ISBN
- 978-3-030-32245-8
- 9783030322441
- DOI
- 10.1007/978-3-030-32245-8_40
- language
- English
- LU publication?
- yes
- id
- f1d90480-d6df-4b9a-b70d-e4f7f86f9790
- date added to LUP
- 2019-12-16 16:21:57
- date last changed
- 2024-10-02 18:32:57
@inproceedings{f1d90480-d6df-4b9a-b70d-e4f7f86f9790, abstract = {{<p>The procedure of aligning a positron emission tomography (PET) image with a common coordinate system, spatial normalization, typically demands a corresponding structural magnetic resonance (MR) image. However, MR imaging is not always available or feasible for the subject, which calls for enabling spatial normalization without MR, MR-less spatial normalization. In this work, we propose a template-free approach to MR-less spatial normalization for [18F]flortaucipir tau PET images. We use a deep neural network that estimates an aligning transformation from the PET input image, and outputs the spatially normalized image as well as the parameterized transformation. In order to do so, the proposed network iteratively estimates a set of rigid and affine transformations by means of convolutional neural network regressors as well as spatial transformer layers. The network is trained and validated on 199 tau PET volumes with corresponding ground truth transformations, and tested on two different datasets. The proposed method shows competitive performance in terms of registration accuracy as well as speed, and compares favourably to previously published results.</p>}}, author = {{Alvén, Jennifer and Heurling, Kerstin and Smith, Ruben and Strandberg, Olof and Schöll, Michael and Hansson, Oskar and Kahl, Fredrik}}, booktitle = {{Medical Image Computing and Computer Assisted Intervention : MICCAI 2019 - 22nd International Conference, Proceedings}}, editor = {{Shen, Dinggang and Yap, Pew-Thian and Liu, Tianming and Peters, Terry M. and Khan, Ali and Staib, Lawrence H. and Essert, Caroline and Zhou, Sean}}, isbn = {{978-3-030-32245-8}}, issn = {{1611-3349}}, language = {{eng}}, month = {{10}}, pages = {{355--363}}, publisher = {{Springer Nature}}, series = {{Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}}, title = {{A Deep Learning Approach to MR-less Spatial Normalization for Tau PET Images}}, url = {{http://dx.doi.org/10.1007/978-3-030-32245-8_40}}, doi = {{10.1007/978-3-030-32245-8_40}}, volume = {{11765 LNCS}}, year = {{2019}}, }