HQColon
收藏资源简介:
HQColon数据集是由哥本哈根大学计算机科学系创建的高分辨率结肠标注数据集。该数据集通过使用基于规则的传统分割方法和交互式机器学习相结合的半自动化管道生成,旨在为nnU-Net模型提供训练数据,以实现从CT结肠造影图像中全自动分割出高分辨率的结肠结构。数据集包含了825名50岁及以上患者的3451个CT扫描,经过筛选和专家验证,最终形成了包含435个高质量标注的CT扫描数据集。该数据集可应用于临床和研究领域,如数字孪生和个性化医疗,有助于推进结肠分割技术的准确性。
The HQColon dataset is a high-resolution colon annotation dataset developed by the Department of Computer Science, University of Copenhagen. It was constructed via a semi-automated pipeline combining rule-based traditional segmentation methods and interactive machine learning, with the goal of providing training data for the nnU-Net model to enable fully automatic segmentation of high-resolution colon structures from CT colonography images. The dataset initially contains 3451 CT scans from 825 patients aged 50 years or older. After screening and expert validation, a final dataset consisting of 435 high-quality annotated CT scans was obtained. This dataset can be applied in clinical and research fields such as digital twins and personalized medicine, helping to advance the accuracy of colon segmentation technologies.

- 1HQColon: A Hybrid Interactive Machine Learning Pipeline for High Quality Colon Labeling and Segmentation哥本哈根大学 · 2025年



