用于心电疾病诊断的深度学习模型库

最新推荐文章于 2024-03-12 15:59:23 发布

处女座程序员的朋友

最新推荐文章于 2024-03-12 15:59:23 发布

阅读量947

点赞数 3

本文链接：https://blog.csdn.net/sinat_22510827/article/details/111598110

版权

该博客介绍了深度学习模型库，用于心电疾病诊断，包括1D-ResNet等7种基线模型，F1分数在0.83-0.90之间。内容涵盖了数据集获取、环境搭建、模型选择，并特别讨论了LSTM在ECG分类中的应用，提供相关代码资源。

摘要由CSDN通过智能技术生成

用于心电疾病诊断的深度学习模型库

github: https://github.com/hzzhangqf0558/ECG_Nets
Baseline model collection of deep learning applied into ECGs. Those baseline models include 1D-ResNet, 1D-DenseNet, 1D-SE_ResNet, 1D-ResNext,1D-SE_ResNetV2, 1D-SE_ResNext and 1D-Top1Net(the champion model in Tianchi competition).

The F1-scores of those baseline models are in range of 0.83-0.90. you could fine-tune the parameters to reach a better level.

Step 1:
downloading datasets. The datasets from Tianchi competition includes dataset A and dataset B. The dataset A consists of 24106 ECGs, of which each has 8 leading records(I，II，V1，V2，V3，V4，V5, V6). while the dataset B consists of 20036 ECGs. Dataset can be downloaded from website.
https://pan.baidu.com/s/1fmCuV5i9oifnUNOsFhV0sA pwd: 8hs2