Communication Dans Un Congrès Année : 2025

Unsupervised Anomaly Detection for Fetal Brain MRI using Two-Stage Denoising Autoencoder (ω-DAE)

Résumé

Fetal germinal matrix and intraventricular hemorrhage (GMH-IVH) is the most common type of fetal brain hemorrhage and holds significant clinical importance. MRI is pivotal for the early detection of subtle prenatal GMH-IVH lesions, which manifest as hypointense signals in T2-weighted MRI. Deep learning emerges as a powerful tool for detecting such lesions from MRI, but they require large datasets from diseased subjects with ground-truth labels for training. Unfortunately, such datasets are difficult to obtain due to the rarity of GMH-IVH (0.5‰–0.9‰) and the challenges in annotating fetal MRI. Moreover, the low signal-to-noise ratio (SNR) in fetal brain MRI causes existing unsupervised anomaly detection methods to struggle with precise lesion segmentation. To address these challenges, we propose the ω-DAE based on 3D convolutional neural networks (CNNs) for unsupervised diagnosis of GMH-IVH.

Fichier principal
Vignette du fichier
OHBM.pdf (1023.16 Ko) Télécharger le fichier
Origine Fichiers produits par l'(les) auteur(s)
Licence

Dates et versions

hal-04974207 , version 1 (03-03-2025)

Licence

Identifiants

  • HAL Id : hal-04974207 , version 1

Citer

Yingqi Hao, Mingxuan Liu, Juncheng Zhu, Hongjia Yang, Yi Liao, et al.. Unsupervised Anomaly Detection for Fetal Brain MRI using Two-Stage Denoising Autoencoder (ω-DAE). 2025 OHBM Annual Meeting, Jun 2025, Brisbane, Australia. ⟨hal-04974207⟩
275 Consultations
206 Téléchargements

Partager

  • More