ECGtizer

ECGtizer presentation

ECGtizer is a tool for digitizing ECGs PDF. We have also work to be able to complete the ECG extracted. Indeed the extracted leads are often represented only on 2.5sec or 5sec whearas the record last 10sec. We have developped a model to complete the leads.

Installation

Ensure you have the required dependencies installed by creating a Conda environment using the provided ecgtizer_env.yml file.

conda env create -f ecgtizer_env.yml
conda activate ecgtizer

How ECGtizer works

ECGtizer create a object from a pdf file:

How to use ECGtizer with bash

python ECGtyzer_main.py "data/PTB-XL/PDF/00121_hr.pdf" 500 "fragmented" --verbose "data/PTB-XL/Digitized/test.xml"

How to use ECGtizer with Jupyter Notebook

1/ First Step: Extraction

from ecgtizer import ECGtizer
ecg_extracted = ECGtizer (path, DPI = 500,extraction_method="fragmented",  verbose = False, DEBUG = False)

path : Location of the PDF file
extraction_method: Method of extraction ("full", "lazy", "fragemented")
DPI : Dot per inch i.e. resolution of the image
verbose : Indicates the different stages of digitisation and their duration
DEBUG : Plot image to show where error can occure\

From this object you can have access to the extracted lead with:

ecg_extracted.extracted_lead

2/ Second Step: Plot

You can plot the extracted leads with the function plot:

ecg_extracted.plot(lead = "",  begin = 0, end = 'inf', save = False)

lead (optional) : The lead you want to plot, if you do not indicate any lead the function will plot all the leads
begin (optional) : The position on the lead you want to start with
end (optional) : The position on the lead you want to end with
completion(optional, bool): If you want to plot the complete lead, save (optinal) : The location to save the plot, if you do not indicate any location the plot will not be saved\

3/ Third Step: Save

You can save the extracted leads into an xml format with the function save_xml:

extracted_lead.save_xml(save)

save: The location to save the xml

4/ Fourth Step: Completion

You can complete the missing information of each leads with the function completion:

ecg_extracted.completion(model_path, device)

model_path (str): location of the model to use for the completion in .pth format
device (torch.device): device to use for the completion ❗❗❗️ The Completed leads will replace the extracted leads you can plot and save the completed leads like you do with the extracted leads ❗❗❗️