VinDr-RibCXR
收藏资源简介:
VinDr-RibCXR是由Vingroup大数据研究所医学影像部创建的一个基准数据集,专门用于胸部X光片中个体肋骨的自动分割和标记。该数据集包含245张胸部X光片,每张图片均由专家手动进行像素级别的肋骨分割和标注。数据集的创建过程涉及使用自家的网络标注工具,确保了标注的准确性和一致性。VinDr-RibCXR数据集的应用领域主要集中在提高深度学习算法在医学影像处理中的性能,特别是在肋骨分割和标记方面,旨在解决传统方法在处理前肋时遇到的困难。
VinDr-RibCXR is a benchmark dataset created by the Medical Imaging Department of Vingroup Big Data Institute, specifically designed for automatic rib segmentation and labeling in chest X-ray images. This dataset includes 245 chest X-ray images, each of which has undergone manual pixel-level rib segmentation and labeling by medical experts. The creation of the VinDr-RibCXR dataset involved the use of an in-house web-based annotation tool, which ensured the accuracy and consistency of the annotations. The main application domains of the VinDr-RibCXR dataset focus on enhancing the performance of deep learning algorithms in medical image processing, particularly in rib segmentation and labeling, with the aim of addressing the difficulties encountered by traditional methods when handling anterior ribs.




