HandDx-200: A Multimodal Dataset of Palmar and Dorsal Hand Images with Biomarkers for Non-Invasive Disease Research

This project, HandDx-200, is a cross-sectional observational study designed to create a publicly available, multimodal dataset for medical AI research. The dataset will combine synchronized RGB and thermal images of the palmar and dorsal sides of both hands with clinical biomarkers, vital signs, and basic demographics from a diverse population. The primary goal is to enable the future development of non-invasive diagnostic tools that use machine learning to detect diseases such as anemia, diabetes mellitus, and cardiovascular conditions. This dataset will serve as a valuable resource for the scientific community, bridging the gap between visual diagnostics and machine learning in preventive medicine.

本项目命名为HandDx-200，是一项横断面观察性研究，旨在构建可公开获取的多模态数据集，用于医学人工智能（Medical AI）研究。该数据集将整合来自多样化研究人群的、同步采集的双手掌侧与背侧RGB图像及热成像图像，以及临床生物标志物、生命体征与基础人口统计学信息。其核心目标是助力未来开发基于机器学习的非侵入式诊断工具，用于检测贫血（Anemia）、糖尿病（Diabetes Mellitus）以及心血管疾病（Cardiovascular Conditions）等病症。本数据集将为科学界提供宝贵的研究资源，填补预防医学领域视觉诊断与机器学习交叉融合的研究空白。