健康数据集在研究和医学教育中起着关键作用,但创建反映真实世界情况的数据集具有挑战性。例如,皮肤病理表现会因外观、严重程度和肤色不同而呈现差异。然而现有的皮肤病学图像数据集往往缺乏对日常病症(如皮疹、过敏和感染)的覆盖,且倾向于浅肤色样本。此外,种族和民族信息经常缺失,这限制了我们评估差异或创建解决方案的能力。
pifu维度 | 2D |
模态 | dermoscopy |
任务类型 | classification |
解剖结构 | 皮肤 |
解剖区域 | 皮肤 |
数据量 | 10408 |
文件格式 | PNG |
SCIN
└── dataset/
├── images
├── scin_app_questions.csv
├── scin_cases.csv
├── scin_label_questions.csv
├── scin_labels.csv
@misc{ward2024crowdsourcing,
title={Crowdsourcing Dermatology Images with Google Search Ads: Creating a Real-World Skin Condition Dataset},
author={Abbi Ward and Jimmy Li and Julie Wang and Sriram Lakshminarasimhan and Ashley Carrick and Bilson Campana and Jay Hartford and Pradeep Kumar S and Tiya Tiyasirichokchai and Sunny Virmani and Renee Wong and Yossi Matias and Greg S. Corrado and Dale R. Webster and Dawn Siegel and Steven Lin and Justin Ko and Alan Karthikesalingam and Christopher Semturs and Pooja Rao},
year={2024},
eprint={2402.18545},
archivePrefix={arXiv},
primaryClass={cs.CY}
}