Deep Learning-based Bone Structure Analyses in Total Hip Arthroplasty Dataset

本数据集名为‘基于深度学习的全髋关节置换术中骨骼结构分析数据集’，由中国科学技术大学创建。数据集包含363个计算机断层扫描（CT）扫描，用于支持深度学习技术在骨骼结构分析中的应用。数据集内容丰富，涵盖了来自不同患者的多样性CT扫描，旨在通过提供高质量的标注数据，解决深度学习在骨骼结构分析中对高质量标注数据的依赖问题。数据集创建过程中采用了非学习基础的骨骼提取和髋臼与股骨头分割等技术，并通过主动学习基础的标注细化过程进一步提高标注质量。该数据集主要应用于全髋关节置换术的计算机辅助诊断和术前规划，以提高手术的精确性和安全性。

This dataset, named Deep Learning-based Bone Structure Analysis Dataset for Total Hip Arthroplasty, was created by the University of Science and Technology of China. It contains 363 computed tomography (CT) scans to support the application of deep learning technologies in bone structure analysis. The dataset includes diverse CT scans from various patients, aiming to address the heavy reliance of deep learning-based bone structure analysis on high-quality annotated data by providing such high-quality labeled datasets. During the dataset construction, non-learning-based techniques including bone extraction, acetabulum and femoral head segmentation were adopted, and the annotation quality was further enhanced through an active learning-based annotation refinement process. This dataset is primarily used for computer-aided diagnosis and preoperative planning of total hip arthroplasty to improve the accuracy and safety of surgical procedures.