SPDA: Superpixel-based Data Augmentation for Biomedical Image Segmentation

Yizhe Zhang, Lin Yang, Hao Zheng, Peixian Liang, Colleen Mangold, Raquel G. Loreto, David P. Hughes, Danny Z. Chen
Proceedings of The 2nd International Conference on Medical Imaging with Deep Learning, PMLR 102:572-587, 2019.

Abstract

Supervised training a deep neural network aims to “teach” the network to mimic human visual perception that is represented by image-and-label pairs in the training data. Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. The SP images thus obtained are then jointly used with the original training images to train a deep learning model. Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. Additional studies also demonstrate that SPDA can practically reduce the generalization gap.

Cite this Paper

BibTeX
@InProceedings{pmlr-v102-zhang19a, title = {SPDA: Superpixel-based Data Augmentation for Biomedical Image Segmentation}, author = {Zhang, Yizhe and Yang, Lin and Zheng, Hao and Liang, Peixian and Mangold, Colleen and Loreto, Raquel G. and Hughes, David P. and Chen, Danny Z.}, booktitle = {Proceedings of The 2nd International Conference on Medical Imaging with Deep Learning}, pages = {572--587}, year = {2019}, editor = {Cardoso, M. Jorge and Feragen, Aasa and Glocker, Ben and Konukoglu, Ender and Oguz, Ipek and Unal, Gozde and Vercauteren, Tom}, volume = {102}, series = {Proceedings of Machine Learning Research}, month = {08--10 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v102/zhang19a/zhang19a.pdf}, url = {https://proceedings.mlr.press/v102/zhang19a.html}, abstract = {Supervised training a deep neural network aims to “teach” the network to mimic human visual perception that is represented by image-and-label pairs in the training data. Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. The SP images thus obtained are then jointly used with the original training images to train a deep learning model. Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. Additional studies also demonstrate that SPDA can practically reduce the generalization gap.} }
Endnote
%0 Conference Paper %T SPDA: Superpixel-based Data Augmentation for Biomedical Image Segmentation %A Yizhe Zhang %A Lin Yang %A Hao Zheng %A Peixian Liang %A Colleen Mangold %A Raquel G. Loreto %A David P. Hughes %A Danny Z. Chen %B Proceedings of The 2nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2019 %E M. Jorge Cardoso %E Aasa Feragen %E Ben Glocker %E Ender Konukoglu %E Ipek Oguz %E Gozde Unal %E Tom Vercauteren %F pmlr-v102-zhang19a %I PMLR %P 572--587 %U https://proceedings.mlr.press/v102/zhang19a.html %V 102 %X Supervised training a deep neural network aims to “teach” the network to mimic human visual perception that is represented by image-and-label pairs in the training data. Superpixelized (SP) images are visually perceivable to humans, but a conventionally trained deep learning model often performs poorly when working on SP images. To better mimic human visual perception, we think it is desirable for the deep learning model to be able to perceive not only raw images but also SP images. In this paper, we propose a new superpixel-based data augmentation (SPDA) method for training deep learning models for biomedical image segmentation. Our method applies a superpixel generation scheme to all the original training images to generate superpixelized images. The SP images thus obtained are then jointly used with the original training images to train a deep learning model. Our experiments of SPDA on four biomedical image datasets show that SPDA is effective and can consistently improve the performance of state-of-the-art fully convolutional networks for biomedical image segmentation in 2D and 3D images. Additional studies also demonstrate that SPDA can practically reduce the generalization gap.
APA
Zhang, Y., Yang, L., Zheng, H., Liang, P., Mangold, C., Loreto, R.G., Hughes, D.P. & Chen, D.Z.. (2019). SPDA: Superpixel-based Data Augmentation for Biomedical Image Segmentation. Proceedings of The 2nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 102:572-587 Available from https://proceedings.mlr.press/v102/zhang19a.html.

Related Material