Endomapper数据集
收藏资源简介:
Endomapper数据集是由Aragón卫生研究所创建的,包含96个完整内窥镜手术视频,总时长超过24小时。该数据集旨在支持内窥镜视觉同时定位与映射(VSLAM)方法的开发与评估,涵盖了内窥镜校准视频、解剖标志、手术标注、分割、重建等多种元数据和标注。数据集的应用领域包括内窥镜导航辅助、盲区检测、息肉测量等,旨在解决内窥镜手术中的定位与导航问题。
The Endomapper dataset was constructed by the Aragón Health Research Institute. It comprises 96 full-length endoscopic surgical videos with a total duration exceeding 24 hours. This dataset is intended to support the development and evaluation of endoscopic Visual Simultaneous Localization and Mapping (VSLAM) approaches, and includes a comprehensive range of metadata and annotations covering endoscopic calibration videos, anatomical landmarks, surgical annotations, segmentation datasets, and 3D reconstruction annotations. Its application areas include endoscopic navigation assistance, blind zone detection, polyp measurement, and other relevant tasks, with the core objective of addressing localization and navigation challenges in endoscopic surgical procedures.




