ATOM

ATOM数据集是由卡内基梅隆大学创建的，旨在通过深度学习解决多模态生物医学图像注册问题。该数据集包含从不同机构收集的3D医学体积数据，如脑部MRI图像和胸部CT等。通过随机选择不同视角的2D切片，ATOM确保了数据的真实性和多样性。数据集的创建过程自动化，无需手动标注，适用于训练智能医学图像注册代理。ATOM的应用领域包括提高疾病预测的准确性和理解非侵入性放射学研究的基础，从而推动医学图像处理和分析的发展。

The ATOM dataset was developed by Carnegie Mellon University, designed to solve multimodal biomedical image registration tasks using deep learning. This dataset includes 3D medical volumetric data collected from diverse institutions, such as brain MRI scans and chest CT scans. By randomly selecting 2D slices from different viewpoints, ATOM ensures the authenticity and diversity of the dataset. The entire creation process of the dataset is fully automated without requiring manual annotation, making it suitable for training intelligent medical image registration AI Agents. The application domains of ATOM include enhancing the accuracy of disease prediction and elucidating the fundamental principles of non-invasive radiological research, thereby advancing the development of medical image processing and analysis.