Aneumo合成动脉瘤血流动力学集
Aneumo 可用于动脉瘤血流动力学代理模型、形态-流场关系学习和快速风险评估研究，尤其适合需要大量三维模拟样本的深度学习方法。
基本信息
资源简介
Aneumo 是面向颅内动脉瘤血流动力学的大规模合成数据集，提供几何和 CFD 相关模拟数据。
下载信息
注册下载
Tips: 该数据集需要在对应的数据源网站注册通过后，才能进行数据下载，注册有对应要求，或者需要收费。
暂未开放公开下载
Tips: 该数据集属于公开下载，应该可以免费公开下载。
免登录有偿下载
Tips: 该数据集 Qianfanghub 可以协助提供有偿下载服务，注意，服务不针对数据相关产权，只是技术服务费。
提供高速下载与技术交付服务（收技术服务费，非数据销售）
暂未开放千方医数集，医疗数据集部分，是为社区服务的公开医疗数据集搜索引擎，并不存储或者下载原始的任何数据。 如果您有其他医疗数据需求，可以和客服联系，或者下工单。我们有强大的三甲医疗机构帮助您提供个性化的医疗数据定制、采集、标注服务。
使用方式
数据集获取
git clone https://github.com/Xigui-Li/Aneumo.git
curl -L -o repo.zip https://github.com/Xigui-Li/Aneumo/archive/refs/heads/main.zip
unzip repo.zip
源站 README 摘录（使用方式）
Aneumo: A Large-Scale Multimodal Aneurysm Dataset with Computational Fluid Dynamics Simulations and Deep Learning Benchmarks
Project Overview
Aneumo is a large-scale, comprehensive, multimodal cerebral aneurysm hemodynamics dataset designed to advance machine learning and computational fluid dynamics (CFD) research.
This dataset is built on 427 real aneurysm geometries and includes:
- 10,660 high-precision 3D models (synthetic aneurysm evolutions).
- A massive 95,940 total simulation cases, uniquely including:
- 85,280 steady-state simulations (under 8 flow conditions).
- 10,660 high-fidelity transient (pulsatile) simulations.
- Pre-computed WSS: Wall Shear Stress (WSS), velocity, and pressure fields are pre-computed for all 95,940 cases.
- Multimodal Inputs: Rich geometric representations including Segmentation Masks, Point Clouds, and Meshes.
<div align=“center”>
This project provides deep learning benchmark code for:
-
Steady-State Models — “Syn-to-Real” task (training on synthetic Aneumo, testing on real AneuX data):
- DeepONet-based model
- Hybrid model combining Swin Transformer with DeepONet (for multimodal inputs: masks + point clouds)
-
Transient Models — Temporal WSS prediction from pulsatile CFD data:
- Aneumo (Temporal DeepONet V2): History encoder (Transformer/MLP) + Swin Transformer geometry encoder + DeepONet with cross-attention
- Baselines: FNO, 3D U-Net, MeshGraphNet
- Cross-geometry Generalization: Multi-case training and evaluation with geometry-split and time-split protocols
<div align=“center”>
<img src=“https://github.com/Xigui-Li/Aneumo/blob/main/fig/network.png?raw=true” width=“800px”>
<p><b>Figure 2:</b> Schematic illustration of the DeepONet-SwinT model architecture for predicting aneurysm hemodynamic parameters.</p>
</div>
Dataset Features and Contributions
Key Contributions
- First Large-Scale, Dual-Mode Hemodynamics Dataset: Provides 95,940 total hemodynamic data samples. This is the first dataset to include 10,660 high-fidelity transient (pulsatile) simulations at this scale, filling a critical gap for spatio-temporal AI models.
- Pre-Computed WSS for All Cases: To maximize usability and lower the barrier for AI research, the critical Wall Shear Stress (WSS) fiel
数据加载示例（图像类）
from PIL import Image
import glob, os
files = (glob.glob(os.path.join(path, "**", "*.png"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.jpg"), recursive=True)
+ glob.glob(os.path.join(path, "**", "*.tif"), recursive=True))
print("图像文件数:", len(files))
img = Image.open(files[0]); print("尺寸/模式:", img.size, img.mode)
# torchvision Dataset 方式:
# from torchvision import datasets
# ds = datasets.ImageFolder(path) # 要求 子目录=类别
目录组织与标注格式以源站说明和下载后实际文件为准。
精度瓶颈？数据缺失？
当前公开数据无法满足您的算法精度？千方提供针对 脑动脉瘤（未破裂） 的高质量、多模态真实临床数据定制解决方案。
