Semantic Segmentation of 3D Medical Images Through a Kaleidoscope: Data from the Osteoarthritis Initiative

Boyeong Woo, Marlon Bran Lorenzana, Craig Engstrom, William Baresic, Jurgen Fripp, Stuart Crozier, Shekhar S. Chandra
Medical Imaging with Deep Learning, PMLR 227:905-917, 2024.

Abstract

While there have been many studies on using deep learning for medical image analysis, the lack of manually annotated data remains a challenge in training a deep learning model for segmentation of medical images. This work shows how the kaleidoscope transform (KT) can be applied to a 3D convolutional neural network to improve its generalizability when the training set is extremely small. In this study, the KT was applied to a context aggregation network (CAN) for semantic segmentation of anatomical structures in knee MR images. In the proposed model, KAN3D, the input image is rearranged into a batch of downsampled images (KT) before the convolution operations, and then the voxels are rearranged back to their original positions (inverse KT) after the convolution operations to produce the predicted segmentation mask for the input image. Compared to the CAN3D (without the KT), the KAN3D was able to reduce overfitting without data augmentation while maintaining a fast training and inference time. The paper discusses the observed advantages and disadvantages of KAN3D.

Cite this Paper

BibTeX
@InProceedings{pmlr-v227-woo24a, title = {Semantic Segmentation of 3D Medical Images Through a Kaleidoscope: Data from the Osteoarthritis Initiative}, author = {Woo, Boyeong and Lorenzana, Marlon Bran and Engstrom, Craig and Baresic, William and Fripp, Jurgen and Crozier, Stuart and Chandra, Shekhar S.}, booktitle = {Medical Imaging with Deep Learning}, pages = {905--917}, year = {2024}, editor = {Oguz, Ipek and Noble, Jack and Li, Xiaoxiao and Styner, Martin and Baumgartner, Christian and Rusu, Mirabela and Heinmann, Tobias and Kontos, Despina and Landman, Bennett and Dawant, Benoit}, volume = {227}, series = {Proceedings of Machine Learning Research}, month = {10--12 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v227/woo24a/woo24a.pdf}, url = {https://proceedings.mlr.press/v227/woo24a.html}, abstract = {While there have been many studies on using deep learning for medical image analysis, the lack of manually annotated data remains a challenge in training a deep learning model for segmentation of medical images. This work shows how the kaleidoscope transform (KT) can be applied to a 3D convolutional neural network to improve its generalizability when the training set is extremely small. In this study, the KT was applied to a context aggregation network (CAN) for semantic segmentation of anatomical structures in knee MR images. In the proposed model, KAN3D, the input image is rearranged into a batch of downsampled images (KT) before the convolution operations, and then the voxels are rearranged back to their original positions (inverse KT) after the convolution operations to produce the predicted segmentation mask for the input image. Compared to the CAN3D (without the KT), the KAN3D was able to reduce overfitting without data augmentation while maintaining a fast training and inference time. The paper discusses the observed advantages and disadvantages of KAN3D.} }
Endnote
%0 Conference Paper %T Semantic Segmentation of 3D Medical Images Through a Kaleidoscope: Data from the Osteoarthritis Initiative %A Boyeong Woo %A Marlon Bran Lorenzana %A Craig Engstrom %A William Baresic %A Jurgen Fripp %A Stuart Crozier %A Shekhar S. Chandra %B Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ipek Oguz %E Jack Noble %E Xiaoxiao Li %E Martin Styner %E Christian Baumgartner %E Mirabela Rusu %E Tobias Heinmann %E Despina Kontos %E Bennett Landman %E Benoit Dawant %F pmlr-v227-woo24a %I PMLR %P 905--917 %U https://proceedings.mlr.press/v227/woo24a.html %V 227 %X While there have been many studies on using deep learning for medical image analysis, the lack of manually annotated data remains a challenge in training a deep learning model for segmentation of medical images. This work shows how the kaleidoscope transform (KT) can be applied to a 3D convolutional neural network to improve its generalizability when the training set is extremely small. In this study, the KT was applied to a context aggregation network (CAN) for semantic segmentation of anatomical structures in knee MR images. In the proposed model, KAN3D, the input image is rearranged into a batch of downsampled images (KT) before the convolution operations, and then the voxels are rearranged back to their original positions (inverse KT) after the convolution operations to produce the predicted segmentation mask for the input image. Compared to the CAN3D (without the KT), the KAN3D was able to reduce overfitting without data augmentation while maintaining a fast training and inference time. The paper discusses the observed advantages and disadvantages of KAN3D.
APA
Woo, B., Lorenzana, M.B., Engstrom, C., Baresic, W., Fripp, J., Crozier, S. & Chandra, S.S.. (2024). Semantic Segmentation of 3D Medical Images Through a Kaleidoscope: Data from the Osteoarthritis Initiative. Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 227:905-917 Available from https://proceedings.mlr.press/v227/woo24a.html.

Related Material