Type
Conference PaperDate
2021-09-21Preprint Posting Date
2021-07-09Online Publication Date
2021-09-21Print Publication Date
2021Permanent link to this record
http://hdl.handle.net/10754/670288
Metadata
Show full item recordAbstract
The reliability of Deep Learning systems depends on their accuracy but also on their robustness against adversarial perturbations to the input data. Several attacks and defenses have been proposed to improve the performance of Deep Neural Networks under the presence of adversarial noise in the natural image domain. However, robustness in computer-aided diagnosis for volumetric data has only been explored for specific tasks and with limited attacks. We propose a new framework to assess the robustness of general medical image segmentation systems. Our contributions are two-fold: (i) we propose a new benchmark to evaluate robustness in the context of the Medical Segmentation Decathlon (MSD) by extending the recent AutoAttack natural image classification framework to the domain of volumetric data segmentation, and (ii) we present a novel lattice architecture for RObust Generic medical image segmentation (ROG). Our results show that ROG is capable of generalizing across different tasks of the MSD and largely surpasses the state-of-the-art under sophisticated adversarial attacks.Citation
Daza, L., Pérez, J. C., & Arbeláez, P. (2021). Towards Robust General Medical Image Segmentation. Lecture Notes in Computer Science, 3–13. doi:10.1007/978-3-030-87199-4_1Sponsors
We thank Amazon Web Services (AWS) for a computational research grant used for the development of this project.Publisher
Springer International PublishingConference/Event name
Medical Image Computing and Computer Assisted Intervention – MICCAI 2021arXiv
2107.04263Additional Links
https://link.springer.com/10.1007/978-3-030-87199-4_1ae974a485f413a2113503eed53cd6c53
10.1007/978-3-030-87199-4_1