a b/README.md 

  1
  
# IntroECG: A full-process library for deep learning on 12-lead electrocardiograms





  
  

  2
  





  
  

  3
  
This repository is meant to be a useful resource library for getting started with deep learning work using electrocardiograms. <br>





  
  

  4
  





  
  

  5
  
## 1-Waveform Extraction





  
  

  6
  
<img align="right" width="60%" src="./Intro_ECG_ValveNet_Model_Full.png">





  
  

  7
  





  
  

  8
  
Scripts and tutorial for extracting raw ECG waveforms from GE Muse or PDFs of ECGs. It also includes examples of how to display and review your ECG data. 





  
  

  9
  





  
  

  10
  
## 2-Generating Synthetic ECG Data





  
  

  11
  
Generate your own synthetic electrocardiograms. Comes with the ability to alter many different aspects of the waveform to test different hypotheses.





  
  

  12
  





  
  

  13
  
## 3-Preprocessing





  
  

  14
  
Key preprocessing steps for cleaning and normalizing ECG data. 





  
  

  15
  





  
  

  16
  
## 4-Models





  
  

  17
  
Different example models we've built to showcase approaches that work for electrocardiograms, in pytorch and tensorflow/keras.





  
  

  18
  





  
  

  19
  
## 5-Training with Ignite and Optuna





  
  

  20
  
A framework built on PyTorch Ignite using Optuna to allow for rapid experimentation and displaying your results using Tensorboard





  
  

  21
  





  
  

  22
  
## 6-Putting it into Practice: The ValveNet Model





  
  

  23
  
See the notebooks we used for generating our figures and key results on our valvular heart disease model published in JACC.





  
  

  24
  





  
  

  25
  
For more details, see the accompanying paper,





  
  

  26
  





  
  

  27
  
> [**Deep Learning Electrocardiographic Analysis for Detection of Left-Sided Valvular Heart Disease**](https://www.jacc.org/doi/10.1016/j.jacc.2022.05.029?utm_medium=email_newsletter&utm_source=jacc&utm_campaign=toc&utm_content=20220801#mmc1)<br/>





  
  

  28
  
  Pierre Elias, Timothy J. Poterucha, Vijay Rajaram, Luca Matos Moller, Victor Rodriguez, Shreyas Bhave, Rebecca T. Hahn, Geoffrey Tison, Sean A. Abreau, Joshua Barrios, Jessica Nicole Torres, J. Weston Hughes, Marco V. Perez, Joshua Finer, Susheel Kodali, Omar Khalique, Nadira Hamid, Allan Schwartz, Shunichi Homma, Deepa Kumaraiah, David J. Cohen, Mathew S. Maurer, Andrew J. Einstein, Tamim Nazif, Martin B. Leon, and Adler J. Perotte. <b>Journal of the American College of Cardiology</b>, August 1, 2022. https://www.jacc.org/doi/10.1016/j.jacc.2022.05.029?utm_medium=email_newsletter&utm_source=jacc&utm_campaign=toc&utm_content=20220801#mmc1





  
  

  29
  





  
  

  30
  
## Development Team





  
  

  31
  
Lead Developers:<br>





  
  

  32
  
[-Pierre Elias](https://twitter.com/PierreEliasMD)<br>





  
  

  33
  
[-Adler Perotte](https://twitter.com/aperotte)<br>





  
  

  34
  





  
  

  35
  
Contributors:<br>





  
  

  36
  
-Vijay Rajaram<br>





  
  

  37
  
-Shengqing Xia<br>





  
  

  38
  
-Alex Wan<br>





  
  

  39
  
-Junyang Jiang<br>





  
  

  40
  
-Yuge Shen<br>





  
  

  41
  
-Han Wang<br>