Angelou0516/DAP_Atlas
收藏资源简介:
DAP Atlas是一个全身CT扫描数据集,包含533个CT扫描体积,每个体积都有142个解剖结构的标注。数据集来源于AutoPET FDG-PET-CT-Lesions队列,分割掩码由Jaus等人(2023年)通过知识聚合、nnU-Net伪标记和基于解剖学指南的后处理生成。数据集包含CT图像和对应的分割掩码,文件格式为NIfTI,数据集大小为约55GB,许可证为Apache-2.0。数据集的详细信息包括模态(CT)、身体部位(全身)、受试者(482名独特患者)、标签(142个解剖结构)、体积形状(通常为512×512×~390)、间距(约0.8×0.8×2.5mm)等。
DAP Atlas is a whole-body CT dataset with 142 anatomical structures segmented across 533 volumes. The CT images come from the AutoPET FDG-PET-CT-Lesions cohort; the segmentation masks were produced by Jaus et al. (2023) via knowledge aggregation across 14 source datasets, nnU-Net pseudo-labelling, and post-processing using anatomical guidelines. The dataset contains CT images and corresponding segmentation masks in NIfTI format, with a total size of approximately 55GB and licensed under Apache-2.0. Detailed information includes modality (CT), body part (full body), subjects (482 unique patients), labels (142 anatomical structures), volume shape (typically 512×512×~390), spacing (~0.8×0.8×2.5mm), etc.




