Veela Challenge - Vessel Extraction and Extrication for Liver Analysis

TOPRAK T., Yazıcı Z. A., Öksüz İ., Koska I. Ö., Bulucu P., GEZER N. S., ...Daha Fazla

35th IEEE International Workshop on Machine Learning for Signal Processing, MLSP 2025, İstanbul, Türkiye, 31 Ağustos - 03 Eylül 2025, (Tam Metin Bildiri) identifier

  • Yayın Türü: Bildiri / Tam Metin Bildiri
  • Doi Numarası: 10.1109/mlsp62443.2025.11204261
  • Basıldığı Şehir: İstanbul
  • Basıldığı Ülke: Türkiye
  • Anahtar Kelimeler: classification, computed tomography, deep learning, Liver vascular tree, segmentation
  • Dokuz Eylül Üniversitesi Adresli: Evet

Özet

Precise segmentation of liver vasculature remains a critical yet challenging objective in clinical procedures, owing to anatomical complexity, inter-patient variability, and inherent radiological artifacts. We introduce VEELA (Vessel Extraction and Extrication for Liver Analysis) challenge, which presents a dataset of 40 abdominal CTA scans from liver transplant donors, derived from the CHAOS challenge. VEELA features comprehensive annotations of hepatic and portal veins, including peripheral vessels. In this paper, the performance of baseline models and the top three performing submissions are provided. The dataset and trained models are publicly available to advance liver vessel segmentation research through VEELA Synapse website.1https://www.synapse.org/Synapse:syn65471967/wiki/631203.