IntrA

IntrA是一个专为深度学习设计的3D颅内动脉瘤数据集，由东京大学和华南理工大学联合创建。该数据集包含116个经过专家手动标注的动脉瘤3D模型，用于支持深度学习模型在颅内动脉瘤诊断和手术规划中的应用。数据集通过重建患者的2D MRA图像生成，不仅包括完整的血管模型，还有自动生成的血管段和手动标注的动脉瘤段。IntrA数据集的特点是数据量虽小但形态多样，适用于研究颅内动脉瘤的分类和部分分割问题，旨在提高手术决策的准确性和效率。

IntrA is a 3D intracranial aneurysm dataset specifically designed for deep learning, co-developed by The University of Tokyo and South China University of Technology. This dataset contains 116 manually annotated 3D aneurysm models by experts, intended to support the application of deep learning models in intracranial aneurysm diagnosis and surgical planning. Generated by reconstructing 2D MRA images of patients, the dataset includes not only complete vascular models, but also automatically generated vascular segments and manually annotated aneurysm segments. Characterized by small scale yet diverse morphology, the IntrA dataset is suitable for research on intracranial aneurysm classification and partial segmentation tasks, aiming to enhance the accuracy and efficiency of surgical decision-making.

搜集汇总

背景与挑战

背景概述

IntrA是一个专为深度学习设计的3D颅内动脉瘤数据集，由东京大学和华南理工大学联合创建，包含116个专家手动标注的动脉瘤3D模型，基于患者2D MRA图像重建生成。该数据集数据量小但形态多样，适用于颅内动脉瘤的分类和部分分割研究，旨在提高手术决策的准确性和效率。