A multi-institute automated segmentation evaluation on a standard dataset: findings from the international workshop on osteoarthritis imaging segmentation challenge
Osteoarthritis and Cartilage, 28, Supplement 1, S304–S305
Abstract
An abstract is not available in the lab archive. View the paper at its source.
This 2020 conference abstract reports a multi-institute evaluation of automated knee MRI segmentation methods. The challenge compared models on a standardized dataset and found that segmentation metrics were only weakly correlated with cartilage thickness error.
Summary
The authors report a standardized, multi-institute challenge evaluating automated segmentation of knee cartilage and menisci in MRI scans. Six teams submitted convolutional neural network methods, which were assessed using segmentation metrics and cartilage thickness estimates. The abstract reports similar segmentation performance across the methods and a weak correlation between segmentation accuracy metrics and cartilage thickness. These findings describe the challenge evaluation; they do not establish clinical readiness.
Cite this work
A. Desai, F. Caliva, C. Iriondo, N. Khosravan, A. Mortazi, S. Jambawalikar, D. Torigian, J. Ellermann, M. Akcakaya, U. Bagci, R. Tibrewala, I. Flament, M. O'brian, S. Majumdar, M. Perslev, A. Pai, C. Igel, E. Dam, S. Gaj, M. Yang, K. Nakamura, X. Li, C. Deniz, V. Juras, R. Regatte, G. Gold, B. Hargreaves, V. Pedoia, A. Chaudhari (2020). A multi-institute automated segmentation evaluation on a standard dataset: findings from the international workshop on osteoarthritis imaging segmentation challenge. Osteoarthritis and Cartilage, 28, Supplement 1, S304–S305 https://doi.org/10.1016/j.joca.2020.02.477
BibTeX