Syn3DWound：用于三维伤口床分析的合成数据集
计算机视觉与模式识别
2024-03-05 v2
摘要
伤口管理构成重大挑战，尤其对卧床患者和老年人而言。准确的诊断与愈合监测可显著受益于现代图像分析，从而提供准确且精确的伤口测量。尽管已有若干现有技术， expansive 且多样化训练数据集的匮乏仍是构建基于机器学习框架的重要障碍。本文介绍 Syn3DWound，一个具有高保真模拟伤口及 2D 和 3D 标注的开源数据集。我们提出了用于自动化 3D 形态测量分析和 2D/3D 伤口分割的基线方法与基准框架。
引用
@article{arxiv.2311.15836,
title = {Syn3DWound: A Synthetic Dataset for 3D Wound Bed Analysis},
author = {Léo Lebrat and Rodrigo Santa Cruz and Remi Chierchia and Yulia Arzhaeva and Mohammad Ali Armin and Joshua Goldsmith and Jeremy Oorloff and Prithvi Reddy and Chuong Nguyen and Lars Petersson and Michelle Barakat-Johnson and Georgina Luscombe and Clinton Fookes and Olivier Salvado and David Ahmedt-Aristizabal},
journal= {arXiv preprint arXiv:2311.15836},
year = {2024}
}
备注
In the IEEE International Symposium on Biomedical Imaging (ISBI) 2024