Model trained on four classes of ECG signal which are follow as ['COVID-19 Patients', 'Normal Person', 'Myocardial Infarction Patients', 'abnormal heart beats']
Dataset can be downloaded from here.
model folder contains the trained model. It trained on the 1000 images per class. Due to less imbalance dataset, I apply the Data Augmentation on it.
Model Metrics
Accuracy: 0.910000
Precision: 0.910000
Recall: 0.910000
F1 score: 0.910000
Confusion matrix:
[[204 5 0 0]
[ 1 181 0 31]
[ 0 0 194 0]
[ 1 34 0 149]]
How to Run the inference?
- create the virtual env
- pip install -r requirements.txt
- Change the path of img in inference.py
- Run the command python3 run.py