An Automated TW3-RUS Bone Age Assessment Method with Ordinal Regression-Based Determination of Skeletal Maturity
Published in Journal of Digital Imaging, 2023
Abstract
The Tanner-Whitehouse 3 radius-ulna-short bones method assesses skeletal maturity from multiple hand and wrist bones, but manual scoring is time-consuming and requires specialized expertise. This work presents an automated TW3-RUS bone-age assessment method that uses ordinal regression to model the ordered nature of skeletal maturity stages. The method detects and analyzes relevant hand-wrist regions, predicts maturity stages for TW3-RUS components, and converts those predictions into automated skeletal maturity and bone-age estimates. By using ordinal relationships rather than treating stages as unrelated classes, the approach aims to improve consistency and clinical usability for pediatric skeletal maturity assessment.
Recommended citation: Dongxu Zhang, Bowen Liu, Yulin Huang, Yang Yan, Shaowei Li, Jinshui He, Shuyun Zhang, Jun Zhang, Ningshao Xia. (2023). "An Automated TW3-RUS Bone Age Assessment Method with Ordinal Regression-Based Determination of Skeletal Maturity." Journal of Digital Imaging, 36(3), 1001-1015. DOI: 10.1007/s10278-023-00794-0.
Publisher page