Certification of Deep Learning Models for Medical Image Segmentation - Inria - Institut national de recherche en sciences et technologies du numérique Access content directly
Conference Papers Year : 2023

Certification of Deep Learning Models for Medical Image Segmentation


In medical imaging, segmentation models have known a significant improvement in the past decade and are now used daily in clinical practice. However, similar to classification models, segmentation models are affected by adversarial attacks. In a safety-critical field like healthcare, certifying model predictions is of the utmost importance. Randomized smoothing has been introduced lately and provides a framework to certify models and obtain theoretical guarantees. In this paper, we present for the first time a certified segmentation baseline for medical imaging based on randomized smoothing and diffusion models. Our results show that leveraging the power of denoising diffusion probabilistic models helps us overcome the limits of randomized smoothing. We conduct extensive experiments on five public datasets of chest X-rays, skin lesions, and colonoscopies, and empirically show that we are able to maintain high certified Dice scores even for highly perturbed images. Our work represents the first attempt to certify medical image segmentation models, and we aspire for it to set a foundation for future benchmarks in this crucial and largely uncharted area.
Fichier principal
Vignette du fichier
paper2489.pdf (1.69 Mo) Télécharger le fichier
Origin Files produced by the author(s)

Dates and versions

hal-04388114 , version 1 (11-01-2024)



Othmane Laousy, Alexandre Araujo, Guillaume Chassagnon, Nikos Paragios, Marie Pierre Revel, et al.. Certification of Deep Learning Models for Medical Image Segmentation. MICCAI 2023 - 26th International Conference on Medical Image Computing and Computer Assisted Intervention, Oct 2023, Vancouver (BC), Canada. pp.611-621, ⟨10.1007/978-3-031-43901-8_58⟩. ⟨hal-04388114⟩
33 View
26 Download



Gmail Mastodon Facebook X LinkedIn More