Robust Multi-Organ Nucleus Segmentation Using a Locally Rotation Invariant Bispectral U-Net

Valentin Oreiller, Julien Fageot, Vincent Andrearczyk, John O. Prior, Adrien Depeursinge
Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, PMLR 172:929-943, 2022.

Abstract

Locally Rotation Invariant (LRI) operators have shown great potential to robustly identify biomedical textures where discriminative patterns appear at random positions and orientations. We build LRI operators through the local projection of the image on circular harmonics followed by the computation of the bispectrum, which is LRI by design. This formulation allows to avoid the discretization of the orientations and does not require any criterion to locally align the descriptors. This operator is used in a convolutional layer resulting in LRI Convolutional Neural Networks (LRI CNN). To evaluate the relevance of this approach, we use it to segment cellular nuclei in histopathological images. We compare the proposed bispectral LRI layer against a standard convolutional layer in a U-Net architecture. While they perform equally in terms of F-score, the LRI CNN provides more robust segmentation with respect to orientation, even when rotational data augmentation is used. This robustness is essential when the relevant pattern may vary in orientation, which is often the case in medical images.

Cite this Paper

BibTeX
@InProceedings{pmlr-v172-oreiller22a, title = {Robust Multi-Organ Nucleus Segmentation Using a Locally Rotation Invariant Bispectral U-Net}, author = {Oreiller, Valentin and Fageot, Julien and Andrearczyk, Vincent and Prior, John O. and Depeursinge, Adrien}, booktitle = {Proceedings of The 5th International Conference on Medical Imaging with Deep Learning}, pages = {929--943}, year = {2022}, editor = {Konukoglu, Ender and Menze, Bjoern and Venkataraman, Archana and Baumgartner, Christian and Dou, Qi and Albarqouni, Shadi}, volume = {172}, series = {Proceedings of Machine Learning Research}, month = {06--08 Jul}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v172/oreiller22a/oreiller22a.pdf}, url = {https://proceedings.mlr.press/v172/oreiller22a.html}, abstract = {Locally Rotation Invariant (LRI) operators have shown great potential to robustly identify biomedical textures where discriminative patterns appear at random positions and orientations. We build LRI operators through the local projection of the image on circular harmonics followed by the computation of the bispectrum, which is LRI by design. This formulation allows to avoid the discretization of the orientations and does not require any criterion to locally align the descriptors. This operator is used in a convolutional layer resulting in LRI Convolutional Neural Networks (LRI CNN). To evaluate the relevance of this approach, we use it to segment cellular nuclei in histopathological images. We compare the proposed bispectral LRI layer against a standard convolutional layer in a U-Net architecture. While they perform equally in terms of F-score, the LRI CNN provides more robust segmentation with respect to orientation, even when rotational data augmentation is used. This robustness is essential when the relevant pattern may vary in orientation, which is often the case in medical images.} }
Endnote
%0 Conference Paper %T Robust Multi-Organ Nucleus Segmentation Using a Locally Rotation Invariant Bispectral U-Net %A Valentin Oreiller %A Julien Fageot %A Vincent Andrearczyk %A John O. Prior %A Adrien Depeursinge %B Proceedings of The 5th International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2022 %E Ender Konukoglu %E Bjoern Menze %E Archana Venkataraman %E Christian Baumgartner %E Qi Dou %E Shadi Albarqouni %F pmlr-v172-oreiller22a %I PMLR %P 929--943 %U https://proceedings.mlr.press/v172/oreiller22a.html %V 172 %X Locally Rotation Invariant (LRI) operators have shown great potential to robustly identify biomedical textures where discriminative patterns appear at random positions and orientations. We build LRI operators through the local projection of the image on circular harmonics followed by the computation of the bispectrum, which is LRI by design. This formulation allows to avoid the discretization of the orientations and does not require any criterion to locally align the descriptors. This operator is used in a convolutional layer resulting in LRI Convolutional Neural Networks (LRI CNN). To evaluate the relevance of this approach, we use it to segment cellular nuclei in histopathological images. We compare the proposed bispectral LRI layer against a standard convolutional layer in a U-Net architecture. While they perform equally in terms of F-score, the LRI CNN provides more robust segmentation with respect to orientation, even when rotational data augmentation is used. This robustness is essential when the relevant pattern may vary in orientation, which is often the case in medical images.
APA
Oreiller, V., Fageot, J., Andrearczyk, V., Prior, J.O. & Depeursinge, A.. (2022). Robust Multi-Organ Nucleus Segmentation Using a Locally Rotation Invariant Bispectral U-Net. Proceedings of The 5th International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 172:929-943 Available from https://proceedings.mlr.press/v172/oreiller22a.html.

Related Material