MultiOrg
收藏资源简介:
MultiOrg数据集是由亥姆霍兹慕尼黑计算健康中心创建的一个用于器官检测任务的多评测者数据集,包含超过400张高分辨率的2D显微镜图像和超过60,000个器官的注释。数据集内容涵盖了从不同生物研究设置中获得的肺器官图像,具有两种类型的器官在不同条件下生长。数据集的创建过程包括图像采集、分割、拼接和注释,由两位专家在不同时间点独立注释,形成三个标签集。该数据集主要应用于生物医学图像分析领域,旨在解决器官检测中的不确定性量化问题,推动深度学习模型在器官检测中的应用。
The MultiOrg dataset is a multi-annotator dataset for organ detection tasks, developed by the Helmholtz Munich Computational Health Center. It contains over 400 high-resolution 2D microscopy images and more than 60,000 organ annotations. The dataset covers lung organ images obtained from various biological research setups, with two types of organs cultured under different conditions. The dataset construction workflow encompasses image acquisition, segmentation, stitching, and annotation: two experts independently conducted annotations at distinct time points, yielding three independent label sets. Primarily utilized in the field of biomedical image analysis, the MultiOrg dataset is designed to tackle the issue of uncertainty quantification in organ detection and advance the deployment of deep learning models for organ detection applications.

- 1MultiOrg: A Multi-rater Organoid-detection Dataset亥姆霍兹慕尼黑计算健康中心 · 2024年



