In-depth Benchmarking of Deep Neural Network Architectures for ECG Diagnosis.

Code used for In-depth Benchmarking of Deep Neural Network Architectures for ECG Diagnosis.

(Part of our code is adopted from Deep Learning for ECG Analysis: Benchmarks and Insights from PTB-XL.)

Setup

pip install poetry (Use python 3.8 or later)

After installing poetry, execute poetry install

Data preparation

Download

cd preparation
./get_dataset.sh

This scripts downloads and stores PTB-XL and CPSC dataset at ./data.

For G12EC dataset, manually download from link, place at ./data/G12EC and unzip the data.

Preparation

poetry run python ptbxl.py
poetry run python cpsc.py
poetry run python g12ec.py

Experiment

cd experiment

For grid search: poetry run python experiment_0.py 0

For multi-label classification: poetry run python experiment_1.py 0

For multi-class classification: poetry run python experiment_2.py 0