Datasets
Models
Downloads: 1
308 lines (307 with data), 12.4 kB
{
"cells": [
{
"cell_type": "code",
"execution_count": 1,
"id": "4c9d40b1",
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import numpy as np\n",
"import cv2\n",
"import matplotlib.pyplot as plt\n",
"import os.path, sys, re\n",
"import time\n",
"from PIL import Image"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "92e8df2f",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"/home/moise/Desktop/Data_Science/Erdos_Institute/ecg-proj/ecg-copy\n"
]
}
],
"source": [
"cd ~/Desktop/Data_Science/Erdos_Institute/ecg-proj/ecg-copy/"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "c5a0e302",
"metadata": {},
"outputs": [],
"source": [
"pathroot = \"data_v1/\""
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "3c9b391a",
"metadata": {},
"outputs": [],
"source": [
"\"\"\"\n",
"Note:\n",
"-----\n",
"\n",
"1)The value for the key 'ECGImagesofMyocardialInfarctionPatients' is a list of dictionaries \n",
"corresponding to the cases\n",
" i) MI <=37\n",
" ii) MI >=38\n",
"2) The value of the key 'ECGImagesofCOVID-19Patients' is broken down as follows:\n",
" i) Binder1_Page file name\n",
" ii) COVID file name\n",
"3) The value of the key 'ECGImagesofPatientthathaveHistoryofMI' corresponds to:\n",
" i) PMI<113\n",
" ii) 113<=PMI<=161 --- ATTENTION: only 12 Leads in the case!\n",
" iii) PMI>=162\n",
"\"\"\"\n",
"superDict = {'NormalPersonECGImages': { 'Lead1':(130,300,640,600), 'Lead2':(641,300,1125,600),'Lead3':(1130,300,1625,600), 'Lead4':(1630,300,2120,600),\\\n",
" 'Lead5':(130,600,640,900), 'Lead6':(641,600,1125,900),'Lead7':(1130,600,1625,900), 'Lead8':(1630,600,2120,900),\\\n",
" 'Lead9':(130,900,640,1200), 'Lead10':(641,900,1125,1200),'Lead11':(1130,900,1625,1200), 'Lead12':(1630,900,2120,1200),\\\n",
" 'Lead13':(130,1205,2120,1450) \n",
" },\n",
" 'ECGImagesofPatientthathaveabnormalheartbeats': { 'Lead1':(130,300,640,600), 'Lead2':(641,300,1125,600),'Lead3':(1130,300,1625,600), 'Lead4':(1630,300,2110,600),\\\n",
" 'Lead5':(130,600,640,900), 'Lead6':(641,600,1125,900),'Lead7':(1130,600,1625,900), 'Lead8':(1630,600,2110,900),\\\n",
" 'Lead9':(130,900,640,1200), 'Lead10':(641,900,1125,1200),'Lead11':(1130,900,1625,1200), 'Lead12':(1630,900,2110,1200),\\\n",
" 'Lead13':(130,1205,2110,1450) \n",
" },\n",
" \n",
" 'ECGImagesofMyocardialInfarctionPatients': [{ 'Lead1':(125,310,625,600), 'Lead2':(640,310,1125,600),'Lead3':(1140,310,1625,600), 'Lead4':(1640,310,2125,600),\\\n",
" 'Lead5':(125,605,625,900), 'Lead6':(640,605,1125,900),'Lead7':(1140,605,1625,900), 'Lead8':(1640,605,2125,900),\\\n",
" 'Lead9':(125,905,625,1200), 'Lead10':(640,905,1125,1200),'Lead11':(1140,905,1625,1200), 'Lead12':(1640,905,2125,1200),\\\n",
" 'Lead13':(125,1205,2125,1495)},\\\n",
" { 'Lead1':(125,300,1125,500), 'Lead2':(1125,300,2125,500),\\\n",
" 'Lead3':(125,500,1125,680), 'Lead4':(1125,500,2125,680),\\\n",
" 'Lead5':(125,680,1125,845), 'Lead6':(1125,680,2125,845),\\\n",
" 'Lead7':(125,845,1125,1000), 'Lead8':(1125,845,2125,1000),\\\n",
" 'Lead9':(125,1000,1125,1150), 'Lead10':(1125,1000,2125,1150),\\\n",
" 'Lead11':(125,1150,1125,1300), 'Lead12':(1125,1150,2125,1300),\\\n",
" 'Lead13':(125,1300,2125,1490)} ],\n",
" \n",
" 'ECGImagesofCOVID-19Patients': [{ 'Lead1':(100,110,300,240), 'Lead2':(300,110,500,240),'Lead3':(500,110,700,240), 'Lead4':(700,110,900,240),\\\n",
" 'Lead5':(100,241,300,350), 'Lead6':(300,241,500,350),'Lead7':(500,241,700,350), 'Lead8':(700,241,900,350),\\\n",
" 'Lead9':(100,351,300,450), 'Lead10':(300,351,500,450),'Lead11':(500,351,700,450), 'Lead12':(700,351,900,450),\\\n",
" 'Lead13':(100,451,900,545)},\\\n",
" { 'Lead1':(125,310,625,600), 'Lead2':(640,310,1125,600),'Lead3':(1140,310,1625,600), 'Lead4':(1640,310,2125,600),\\\n",
" 'Lead5':(125,605,625,900), 'Lead6':(640,605,1125,900),'Lead7':(1140,605,1625,900), 'Lead8':(1640,605,2125,900),\\\n",
" 'Lead9':(125,905,625,1200), 'Lead10':(640,905,1125,1200),'Lead11':(1140,905,1625,1200), 'Lead12':(1640,905,2125,1200),\\\n",
" 'Lead13':(125,1205,2125,1495)} ],\n",
" \n",
" 'ECGImagesofPatientthathaveHistoryofMI': [{ 'Lead1':(125,310,625,600), 'Lead2':(640,310,1125,600),'Lead3':(1140,310,1625,600), 'Lead4':(1640,310,2125,600),\\\n",
" 'Lead5':(125,605,625,900), 'Lead6':(640,605,1125,900),'Lead7':(1140,605,1625,900), 'Lead8':(1640,605,2125,900),\\\n",
" 'Lead9':(125,905,625,1200), 'Lead10':(640,905,1125,1200),'Lead11':(1140,905,1625,1200), 'Lead12':(1640,905,2125,1200),\\\n",
" 'Lead13':(125,1205,2125,1495)},\\\n",
" { 'Lead1':(125,410,1125,590), 'Lead2':(1125,410,2125,590),\\\n",
" 'Lead3':(125,600,1125,815), 'Lead4':(1125,600,2125,815),\\\n",
" 'Lead5':(125,815,1125,1035), 'Lead6':(1125,815,2125,1035),\\\n",
" 'Lead7':(125,1035,1125,1235), 'Lead8':(1125,1035,2125,1235),\\\n",
" 'Lead9':(125,1235,1125,1350), 'Lead10':(1125,1235,2125,1350),\\\n",
" 'Lead11':(125,1350,1125,1490), 'Lead12':(1125,1350,2125,1490),\\\n",
" },\\\n",
" { 'Lead1':(125,290,1125,490), 'Lead2':(1125,290,2125,490),\\\n",
" 'Lead3':(125,490,1125,650), 'Lead4':(1125,490,2125,650),\\\n",
" 'Lead5':(125,650,1125,810), 'Lead6':(1125,650,2125,810),\\\n",
" 'Lead7':(125,810,1125,1000), 'Lead8':(1125,810,2125,1000),\\\n",
" 'Lead9':(125,1000,1125,1190), 'Lead10':(1125,1000,2125,1190),\\\n",
" 'Lead11':(125,1190,1125,1315), 'Lead12':(1125,1190,2125,1315),\\\n",
" 'Lead13':(125,1315,2125,1550)\n",
" }] \n",
"}"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "310cafaf",
"metadata": {},
"outputs": [],
"source": [
"dirList=['NormalPersonECGImages','ECGImagesofPatientthathaveabnormalheartbeats',\n",
" 'ECGImagesofMyocardialInfarctionPatients','ECGImagesofCOVID-19Patients',\n",
" 'ECGImagesofPatientthathaveHistoryofMI']"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "76c9879b",
"metadata": {},
"outputs": [],
"source": [
"def processor(dirs,path2,im,of,crop_dict,numLeads=13,threshold_level=50):\n",
" for i in range(1,numLeads+1):\n",
" imCrop = im.crop(box=crop_dict['Lead'+str(i)])\n",
" target = of+'-Cropped_lead'+str(i)\n",
" imCrop.save(os.path.join(path2,target+'.png'), \"PNG\", quality=100)\n",
" img = cv2.imread(os.path.join(path2,target+'.png'))\n",
" gray = cv2.cvtColor(img, cv2.COLOR_BGR2GRAY)\n",
" coords = np.column_stack(np.where(gray < threshold_level))\n",
" coords[:,[0,1]]=coords[:,[1,0]]\n",
" csvFolder = os.path.join(targetFolder,dirs)\n",
" if not os.path.exists(csvFolder):\n",
" os.makedirs(csvFolder)\n",
" np.savetxt(os.path.join(csvFolder,target+'.csv'),coords) "
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "15742532",
"metadata": {},
"outputs": [],
"source": [
"def crop2csv(dirs,path,path2,dirList,superDict,pattern='.jpg'):\n",
" count = 0\n",
" for item in os.listdir(path):\n",
" if os.path.isfile(os.path.join(path,item)):\n",
" of, oe = os.path.splitext(item)\n",
" if oe == pattern:\n",
" im = Image.open(os.path.join(path,item))\n",
" if dirs == dirList[2]:\n",
" if int(of[3:]) <= 37:\n",
" processor(dirs,path2,im,of,superDict[dirs][0])\n",
" else:\n",
" processor(dirs,path2,im,of,superDict[dirs][1])\n",
" elif dirs == dirList[3]:\n",
" if of[0:5] == 'COVID':\n",
" processor(dirs,path2,im,of,superDict[dirs][1])\n",
" else:\n",
" processor(dirs,path2,im,of,superDict[dirs][0])\n",
" elif dirs == dirList[4]:\n",
" if int(of[4:]) < 113:\n",
" processor(dirs,path2,im,of,superDict[dirs][0])\n",
" elif (int(of[4:]) >= 113) and (int(of[4:]) <= 161) :\n",
" processor(dirs,path2,im,of,superDict[dirs][1],numLeads=12)\n",
" else:\n",
" processor(dirs,path2,im,of,superDict[dirs][2])\n",
" else:\n",
" processor(dirs,path2,im,of,superDict[dirs])\n",
" count = count+1\n",
" return count"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "61da0528",
"metadata": {
"scrolled": true
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Processing ECGImagesofPatientthathaveabnormalheartbeats folder ...\n",
"546 files processed in this folder in 229 sec...\n",
"\n",
"Processing ECGImagesofPatientthathaveHistoryofMI folder ...\n",
"203 files processed in this folder in 81 sec...\n",
"\n",
"Processing ECGImagesofCOVID-19Patients folder ...\n",
"250 files processed in this folder in 27 sec...\n",
"\n",
"Processing NormalPersonECGImages folder ...\n",
"859 files processed in this folder in 348 sec...\n",
"\n",
"Processing ECGImagesofMyocardialInfarctionPatients folder ...\n",
"74 files processed in this folder in 29 sec...\n",
"\n"
]
}
],
"source": [
"## Cropping\n",
"targetFolder = 'CSV_data_v1'\n",
"if not os.path.exists(targetFolder):\n",
" os.makedirs(targetFolder)\n",
"t0 = time.time()\n",
"for dirs in os.listdir(pathroot):\n",
" t = time.time()\n",
" print('Processing {0} folder ...'.format(dirs))\n",
" if not os.path.isfile(dirs):\n",
" path = os.path.join(pathroot,dirs)\n",
" path2 = os.path.join(path,\"Cropped_Images\")\n",
" if not os.path.exists(path2):\n",
" os.makedirs(path2)\n",
" fileCount = crop2csv(dirs,path,path2,dirList,superDict) \n",
" t = time.time()-t\n",
" print('{0} files processed in this folder in {1} sec...\\n'.format(fileCount,round(t)))\n",
"t0 = time.time()-t0"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "ab0d386d",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"id": "bb1756ac",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"id": "7ff6cfcd",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"id": "5ad56a7b",
"metadata": {},
"outputs": [],
"source": []
},
{
"cell_type": "code",
"execution_count": null,
"id": "f2b319f8",
"metadata": {},
"outputs": [],
"source": []
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.10"
}
},
"nbformat": 4,
"nbformat_minor": 5
}