Brain tumor segmentation using a generative model with an RBM prior on tumor shape
(2016) 1st International Workshop on Brainlesion, Brainles 2015 Held in Conjunction with International Conference on Medical Image Computing for Computer-Assisted Intervention, MICCAI 2015 In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) 9556. p.168-180- Abstract
In this paper, we present a fully automated generative method for brain tumor segmentation in multi-modal magnetic resonance images. The method is based on the type of generative model often used for segmenting healthy brain tissues, where tissues are modeled by Gaussian mixture models combined with a spatial atlas-based tissue prior. We extend this basic model with a tumor prior, which uses convolutional restricted Boltzmann machines (cRBMs) to model the shape of both tumor core and complete tumor, which includes edema and core. The cRBMs are trained on expert segmentations of training images, without the use of the intensity information in the training images. Experiments on public benchmark data of patients suffering from low- and... (More)
In this paper, we present a fully automated generative method for brain tumor segmentation in multi-modal magnetic resonance images. The method is based on the type of generative model often used for segmenting healthy brain tissues, where tissues are modeled by Gaussian mixture models combined with a spatial atlas-based tissue prior. We extend this basic model with a tumor prior, which uses convolutional restricted Boltzmann machines (cRBMs) to model the shape of both tumor core and complete tumor, which includes edema and core. The cRBMs are trained on expert segmentations of training images, without the use of the intensity information in the training images. Experiments on public benchmark data of patients suffering from low- and high-grade gliomas show that the method performs well compared to current state-of-the-art methods, while not being tied to any specific imaging protocol.
(Less)
- author
- Agn, Mikael ; Puonti, Oula ; Rosenschöld, Per Munck Af LU ; Law, Ian and Van Leemput, Koen
- publishing date
- 2016
- type
- Chapter in Book/Report/Conference proceeding
- publication status
- published
- subject
- host publication
- Brainlesion : Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 1st International Workshop, Brainles 2015 Held in Conjunction with MICCAI 2015, Revised Selected Papers - Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 1st International Workshop, Brainles 2015 Held in Conjunction with MICCAI 2015, Revised Selected Papers
- series title
- Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
- editor
- Reyes, Mauricio ; Crimi, Alessandro ; Maier, Oskar ; Maier, Oskar ; Handels, Heinz and Menze, Bjoern
- volume
- 9556
- pages
- 168 - 180
- publisher
- Springer
- conference name
- 1st International Workshop on Brainlesion, Brainles 2015 Held in Conjunction with International Conference on Medical Image Computing for Computer-Assisted Intervention, MICCAI 2015
- conference location
- Munich, Germany
- conference dates
- 2015-10-05 - 2015-10-05
- external identifiers
-
- scopus:84961574865
- ISSN
- 1611-3349
- 0302-9743
- ISBN
- 9783319308579
- DOI
- 10.1007/978-3-319-30858-6_15
- language
- English
- LU publication?
- no
- additional info
- Publisher Copyright: © Springer International Publishing Switzerland 2016.
- id
- 34b43c6f-624c-43af-b047-8257ccaf2cf2
- date added to LUP
- 2023-07-31 13:36:21
- date last changed
- 2024-03-08 04:08:36
@inproceedings{34b43c6f-624c-43af-b047-8257ccaf2cf2, abstract = {{<p>In this paper, we present a fully automated generative method for brain tumor segmentation in multi-modal magnetic resonance images. The method is based on the type of generative model often used for segmenting healthy brain tissues, where tissues are modeled by Gaussian mixture models combined with a spatial atlas-based tissue prior. We extend this basic model with a tumor prior, which uses convolutional restricted Boltzmann machines (cRBMs) to model the shape of both tumor core and complete tumor, which includes edema and core. The cRBMs are trained on expert segmentations of training images, without the use of the intensity information in the training images. Experiments on public benchmark data of patients suffering from low- and high-grade gliomas show that the method performs well compared to current state-of-the-art methods, while not being tied to any specific imaging protocol.</p>}}, author = {{Agn, Mikael and Puonti, Oula and Rosenschöld, Per Munck Af and Law, Ian and Van Leemput, Koen}}, booktitle = {{Brainlesion : Glioma, Multiple Sclerosis, Stroke and Traumatic Brain Injuries - 1st International Workshop, Brainles 2015 Held in Conjunction with MICCAI 2015, Revised Selected Papers}}, editor = {{Reyes, Mauricio and Crimi, Alessandro and Maier, Oskar and Maier, Oskar and Handels, Heinz and Menze, Bjoern}}, isbn = {{9783319308579}}, issn = {{1611-3349}}, language = {{eng}}, pages = {{168--180}}, publisher = {{Springer}}, series = {{Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)}}, title = {{Brain tumor segmentation using a generative model with an RBM prior on tumor shape}}, url = {{http://dx.doi.org/10.1007/978-3-319-30858-6_15}}, doi = {{10.1007/978-3-319-30858-6_15}}, volume = {{9556}}, year = {{2016}}, }