PADDIT: Probabilistic Augmentation of Data using Diffeomorphic Image Transformation
Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Standard
PADDIT : Probabilistic Augmentation of Data using Diffeomorphic Image Transformation. / Orbes-Arteaga, Mauricio; Sørensen, Lauge; Cardoso , Jorge ; Modat , Marc ; Ourselin, Sebastien; Sommer, Stefan Horst; Nielsen, Mads; Igel, Christian; Pai, Akshay Sadananda Uppinakudru.
Proceedings, SPIE 10949, Medical Imaging 2019: Image Processing . SPIE - International Society for Optical Engineering, 2019.Publikation: Bidrag til bog/antologi/rapport › Konferencebidrag i proceedings › Forskning › fagfællebedømt
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - GEN
T1 - PADDIT
AU - Orbes-Arteaga, Mauricio
AU - Sørensen, Lauge
AU - Cardoso , Jorge
AU - Modat , Marc
AU - Ourselin, Sebastien
AU - Sommer, Stefan Horst
AU - Nielsen, Mads
AU - Igel, Christian
AU - Pai, Akshay Sadananda Uppinakudru
PY - 2019
Y1 - 2019
N2 - For proper generalization performance of convolutional neural networks (CNNs) in medical image segmentation, the learnt features should be invariant under particular non-linear shape variations of the input. To induce invariance in CNNs to such transformations, we propose Probabilistic Augmentation of Data using Diffeomorphic Image Transformation (PADDIT) – a systematic framework for generating realistic transformations that can be used to augment data for training CNNs. The main advantage of PADDIT is the ability to produce transformations that capture the morphological variability in the training data. To this end, PADDIT constructs a mean template which represents the main shape tendency of the training data. A Hamiltonian Monte Carlo(HMC) scheme is used to sample transformations which warp the training images to the generated mean template. Augmented images are created by warping the training images using the sampled transformations. We show that CNNs trained with PADDIT outperforms CNNs trained without augmentation and with generic augmentation (0.2 and 0.15 higher dice accuracy respectively) in segmenting white matter hyperintensities from T1 and FLAIR brain MRI scans.
AB - For proper generalization performance of convolutional neural networks (CNNs) in medical image segmentation, the learnt features should be invariant under particular non-linear shape variations of the input. To induce invariance in CNNs to such transformations, we propose Probabilistic Augmentation of Data using Diffeomorphic Image Transformation (PADDIT) – a systematic framework for generating realistic transformations that can be used to augment data for training CNNs. The main advantage of PADDIT is the ability to produce transformations that capture the morphological variability in the training data. To this end, PADDIT constructs a mean template which represents the main shape tendency of the training data. A Hamiltonian Monte Carlo(HMC) scheme is used to sample transformations which warp the training images to the generated mean template. Augmented images are created by warping the training images using the sampled transformations. We show that CNNs trained with PADDIT outperforms CNNs trained without augmentation and with generic augmentation (0.2 and 0.15 higher dice accuracy respectively) in segmenting white matter hyperintensities from T1 and FLAIR brain MRI scans.
U2 - 10.1117/12.2512520
DO - 10.1117/12.2512520
M3 - Article in proceedings
BT - Proceedings, SPIE 10949, Medical Imaging 2019: Image Processing
PB - SPIE - International Society for Optical Engineering
Y2 - 16 February 2019 through 21 February 2019
ER -
ID: 215456614