Szczegóły publikacji
Opis bibliograficzny
Improving segmentation of hypoxic ischemic encephalopathy lesions by heavy data augmentation: contribution to the BONBID challenge / Marek WODZIŃSKI, Henning Müller // W: AI for brain lesion detection and trauma video action recognition : first BONBID-HIE lesion segmentation challenge and first Trauma Thompson Challenge held in conjunction with MICCAI 2023 : Vancouver, BC, Canada, October 16 and 12, 2023 : proceedings / eds. Rina Bao, [et al.]. — Cham : Springer, cop. 2025. — (Lecture Notes in Computer Science ; ISSN 0302-9743 ; 14567). — ISBN: 978-3-031-71625-6; e-ISBN: 978-3-031-71626-3. — S. 28–33. — Bibliogr., Abstr. — Publikacja dostępna online od: 2024-10-24. — M. Wodziński – dod. afiliacja: University of Applied Sciences Western Switzerland (HES-SO Valais), Information Systems Institute, Sierre, Switzerland
Autorzy (2)
- AGHWodziński Marek
- Müller Henning
Dane bibliometryczne
| ID BaDAP | 156963 |
|---|---|
| Data dodania do BaDAP | 2025-01-30 |
| DOI | 10.1007/978-3-031-71626-3_4 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Springer |
| Konferencja | Medical Image Computing and Computer-Assisted Intervention 2023 |
| Czasopismo/seria | Lecture Notes in Computer Science |
Abstract
Hypoxic ischemic encephalopathy is a birth complication strongly affecting infants often resulting in death or disabilities. The underlying pathological events result from incorrect cerebral blood flow and thus complications with oxygen delivery to the brain. An automatic segmentation of hypoxic ischemic encephalopathy lesions is a crucial step in the clinical care. Therefore, to address the problem, a dedicated challenge named BONBID-HIE was organized jointly with the MICCAI 2023 conference. This work presents the contribution of the MedGIFT team to the BONBID-HIE challenge. The main idea behind the proposed method was to improve the deep network generalizability by heavy data augmentation. We show that the heavy data augmentation strongly improves the results compared to the baseline, by more than 0.2 in terms of the Dice score.
