This repository provides PyTorch implementation of deep learning model for estimating age from electrocardiogram (ECG). This was the winning solution for the MAIC ECG AI 2023 challenge.
Create a virtual environment and activate it.
conda create -n ecgai python=3.9
conda activate ecgai
Install required packages
git clone https://github.com/jwc-rad/ecg-ai.git
cd ecg-ai
pip install -r requirements.txt
cd ..
git clone https://github.com/jwc-rad/MISLight.git
cd MISLight
pip install -e .
cd ..ecg-ai
The current configurations assume the input shape as 12x5000
, a 10-second 12-lead ECG with the sampling frequence of 500Hz.
There are some public ECG datasets with the same data format as follows:
python train.py experiment=maic_v1 paths.data_root_dir=${PATH_TO_DATA_DIRECTORY}
- For more options, get started from
config/experiment/maic_v1.yaml