Skip to content

Latest commit

 

History

History
51 lines (36 loc) · 2.25 KB

Test Instructions.md

File metadata and controls

51 lines (36 loc) · 2.25 KB
Raw

Testing the Diagno Web App

  1. Visit Diagno Web App
  2. Download the sample recordings from here. (If you have your own ECG recording device you too can generate JSON Files yourselves. See JSON File Format )
  3. Click 'Choose File'/'Browse' and upload any of the sample recordings.
  4. You'll see 12 plots for the visualization of 12-lead ECG signals.
  5. Now click 'Analyse'. You'll get the machine-generated predictions !!

(OPTIONAL) Generating JSON files yourselves using your own ECG Hardware

JSON File Format

  • 12x4096 JSON Array. 4096 samples per each ECG Lead
  • Signals must be normalized (Each data value between 0 and 1 (Float32) )
  • Sampling rate 400Hz. If the recorded at a different sampling rate, resample to 400Hz

Example:

'[
    [0.08933808901152494, 0.08622472008049079, 0.08748484483946328, ..., 0.08669382432404785, 0.08727538263812228],
    [0.08681173974874626, 0.0872001459641301, 0.08686349999574926, ..., 0.08716319715879413, 0.0868887008560258],
    [0.0843244646650618, 0.08458883624778389, 0.08436451200063877, ..., 0.08456321887653273, 0.08440538068028715],
    [0.07692143752869449, 0.07906505602469656, 0.07945948802502142, ..., 0.0793007000280487, 0.07941703115399325],
    ........................,
    [0.09979861811156633, 0.099748679613499, 0.09982761038949958, ..., 0.09951139681462995, 0.09730531724294086]
    [0.09711884099772175, 0.09732782051344732, 0.09711130487391956,..., 0.0973777123000035, 0.09690846013666106]
]'

Python Functions to Resample and Normalise

import numpy as np
from scipy import signal

def resample(input_signal, Fs_in, Fs_out):
    # input_signal=[[],[],[],...] # 12xL
    n_samples_in = input_signal.shape[1]
    n_samples_out =  round(n_samples_in * Fs_out/Fs_in)
    
    output_signals=[]
    for i in range(12):
        output_signals.append(signal.resample(input_signal[i], n_samples_out))
        
    return np.vstack(output_signals)

def normalize(X):
    Xmin=np.amin(X,keepdims=True,axis=1)
    Xmax=np.amax(X,keepdims=True,axis=1)
    return (X-Xmin)/(Xmax-Xmin)