BAGLS: Images Dataset

This is an Images part of a BAGLS: Benchmark for Automatic Glottis Segmentation dataset - the first large-scale, publicly available dataset of endoscopic high-speed video with frame-wise segmentation annotations. It has been collected in a collaboration of seven institutions and features a total of 59,250 frames. The BAGLS dataset aims to provide a baseline as robust as possible. Therefore, it was created in such a way that it contains diverse samples from a variety of data sources.

此为BAGLS：自动声门分割基准数据集的图像部分——该数据集是首个大规模、公开可用的内窥镜高速视频帧级分割标注数据集。该数据集由七家机构联合收集，包含总计59,250帧图像。BAGLS数据集旨在提供一个尽可能稳健的基准。因此，其构建方式确保了包含来自多种数据源的多样化样本。

搜集汇总

背景与挑战

背景概述

BAGLS数据集是一个包含59,250张内窥镜高速视频帧的医学图像数据集，专门用于声门自动分割的基准测试。该数据集由七个国际机构合作创建，涵盖了多样化的患者群体、设备配置和临床环境，并提供了专家标注的分割掩码，旨在支持深度学习模型的训练和评估。