2020 CATARACTS语义分割挑战赛
图像与视频处理
2022-02-25 v2 计算机视觉与模式识别
摘要
手术场景分割对于解剖结构与器械定位至关重要，并可进一步用于评估手术过程中的组织-器械交互。2017年，白内障手术自动器械标注挑战赛（CATARACTS）发布了50个白内障手术视频及配套的器械使用标注。这些标注包含帧级器械存在信息。2020年，我们从CATARACTS训练集中25个视频采样得到4670幅图像，并发布了针对解剖结构与器械的像素级语义标注。2020 CATARACTS语义分割挑战赛作为2020 MICCAI内镜视觉（EndoVis）挑战赛的子挑战赛，提出了三个子任务以评估参赛方案在解剖结构与器械分割上的表现。其性能在一个隐藏测试集上评估，该测试集包含来自CATARACTS测试集10个视频的531幅图像。
引用
@article{arxiv.2110.10965,
title = {2020 CATARACTS Semantic Segmentation Challenge},
author = {Imanol Luengo and Maria Grammatikopoulou and Rahim Mohammadi and Chris Walsh and Chinedu Innocent Nwoye and Deepak Alapatt and Nicolas Padoy and Zhen-Liang Ni and Chen-Chen Fan and Gui-Bin Bian and Zeng-Guang Hou and Heonjin Ha and Jiacheng Wang and Haojie Wang and Dong Guo and Lu Wang and Guotai Wang and Mobarakol Islam and Bharat Giddwani and Ren Hongliang and Theodoros Pissas and Claudio Ravasio and Martin Huber and Jeremy Birch and Joan M. Nunez Do Rio and Lyndon da Cruz and Christos Bergeles and Hongyu Chen and Fucang Jia and Nikhil KumarTomar and Debesh Jha and Michael A. Riegler and Pal Halvorsen and Sophia Bano and Uddhav Vaghela and Jianyuan Hong and Haili Ye and Feihong Huang and Da-Han Wang and Danail Stoyanov},
journal= {arXiv preprint arXiv:2110.10965},
year = {2022}
}