2026-04-17T02:34:55Zhttps://recercat.cat/oai/request

oai:recercat.cat:2445/2199642025-12-05T09:54:35Zcom_2072_1057col_2072_478917col_2072_478920

Leveraging epistemic uncertainty to improve tumour segmentation in breast MRI: an exploratory analysis Joshi, Smriti Osuala, Richard Garrucho, Lidia Tsirikoglou, Apostolia Riego, Javier del Gwoździewicz, Katarzyna Kushibar, Kaisar Díaz, Oliver Lekadir, Karim, 1977- Imatges mèdiques Aprenentatge automàtic Càncer de mama Imaging systems in medicine Machine learning Breast cancer Medical image segmentation has improved with deep-learning methods, especially for tumor segmentation. However, variability in tumor shapes, sizes, and enhancement remains a challenge. Breast MRI adds further uncertainty due to anatomical differences. Informing clinicians about result reliability and using model uncertainty to improve predictions are essential. We study Monte-Carlo Dropout for generating multiple predictions and finding consensus segmentation. Our approach reduces false positives using per-pixel uncertainty and improves segmentation metrics. In addition, we study the correlation of model performance to the perceived ease of manual segmentation. Finally, we compare the per-pixel uncertainty with the inter-rater variability as segmented by six different radiologists. Our code is available at https://github.com/smriti-joshi/uncertainty-segmentation-mcdropout.git. 2025-03-25T08:07:38Z 2025-03-25T08:07:38Z 2024 info:eu-repo/semantics/conferenceObject info:eu-repo/semantics/acceptedVersion https://hdl.handle.net/2445/219964 eng Versió postprint de la comunicació publicada a: https://doi.org/10.1117/12.3006783 Comunicació a: Proc. SPIE 12926, Medical Imaging 2024: Image Processing, 1292616 (2 April 2024) Proceedings SPIE 12926 https://doi.org/10.1117/12.3006783 (c) SPIE, 2024 info:eu-repo/semantics/openAccess 9 p. application/pdf SPIE Comunicacions a congressos (Matemàtiques i Informàtica)