def get_data_for_patient(patient):
signal, fields = wfdb.rdsamp(f'{DATA_DIR}/data/{patient}')
fs = fields['fs']
r_peaks_ix = processing.gqrs_detect(signal[:, 0], fs=fs, threshold=1.0)
features = feature_selection(signal, fs, r_peaks_ix)
return features
def get_rr_peaks_indices(record, max_bpm=230):
"""
:param record_name:
:param database:
:param max_bpm:
:return: list of timestamps for
"""
qrs_inds = processing.gqrs_detect(sig=record.p_signal[:, 0],
fs=record.fs)
search_radius = int(record.fs * 60 / max_bpm)
corrected_peak_inds = processing.correct_peaks(
record.p_signal[:, 0], peak_inds=qrs_inds,
search_radius=search_radius, smooth_window_size=150)
result = []
discarded_count = 0
prev_pi = -1
for pi in corrected_peak_inds:
if pi != prev_pi:
result.append(pi)
else:
discarded_count += 1
prev_pi = pi
# returns r-r peaks timestamps
result = np.array(result) / record.fs
return result, discarded_count
def test_pure_prediction(self):
record_name = self.mitdb + "100"
sig, fields = wfdb.rdsamp(record_name, channels=[0])
res = processing.gqrs_detect(sig, fs=fields['fs'])
wfdb.wrann("100",
'atr',
res,
write_dir="data",
symbol=(['N'] * len(res)))
print(res)
self.assertIsNotNone(res)
def getQRSLocations(file_path):
'''
get numpy list of QRS Locations
Returns:
qrs_locs (numpy list): list of QRS locations in the signal
'''
record = wfdb.rdrecord(file_path, channels=[0])
qrs_locs = processing.gqrs_detect(record.p_signal[:, 0], fs=record.fs)
return qrs_locs
def filter_ecg(ecg, fs, get_heart_rate=False, cutoff_frequencies=[0.5, 100.0]):
nyq_f = fs / 2
w_low = cutoff_frequencies[0] / (math.pi * nyq_f)
w_high = cutoff_frequencies[1] / (math.pi * nyq_f)
filt_order = 4
b, a = scipy.signal.butter(filt_order, [w_low, w_high], btype='bandpass')
processed_ecg = scipy.signal.filtfilt(b, a, ecg)
if get_heart_rate:
qrs_inds = processing.gqrs_detect(sig=ecg, fs=fs)
hr = processing.compute_hr(sig_len=len(ecg), qrs_inds=qrs_inds, fs=fs)
return (processed_ecg, hr)
else:
return (processed_ecg)
def get_data(age_start, age_stop, limit=1):
"""
Получение данных пациентов в промежутке возраста от age_start до age_stop (limit по количеству пациентов)
:return: dict
"""
data_for_age_range = df.loc[(age_start < df['Age']) & (df['Age'] < age_stop)]
result = {'data': []}
for i in range(0, limit):
try:
pdf = data_for_age_range.iloc[i]
fid = pdf['ID']
signal, fields = wfdb.rdsamp(f'{DATA_DIR}/data/{fid}')
qrs_locs = processing.gqrs_detect(signal[:, 0], fs=fields['fs'])
# result['data'].append({'ecg_data': signal.tolist(), 'qrs_ix': qrs_locs.tolist(), 'fields': fields})
patient_data = {
'avatar': f'require("~/assets/images/users/avatar-{randint(1, 4)}.jpg")',
'name': choice(["Смирнов", "Кузнецов", "Попов", "Васильев", "Петров", "Соколов", "Михайлов",
"Фролов", "Журавлёв", "Николаев", "Крылов", "Максимов", "Сидоров", "Осипов", "Белоусов",
"Федотов", "Дорофеев", "Егоров", "Матвеев", "Бобров", "Дмитриев","Калинин", 'Петров', "Васильев", "Тылык", "Иванов", "Пушкин", "Лермонтов", "Данилин", "Ванилин", "Герц", "Тесла", "Улец", "Кершин", "Ким"])+' '+choice(['И.', "А.", "В.", "Д.", "К."])+choice(['И.', "А.", "В.", "Д."]),
'age': str(pdf['Age']),
'id': str(pdf['ID']),
'diagnose': pdf['Rhythms'],
'state': 'Отклонение от нормы' if pdf['Rhythms'] != 'Sinus rhythm' else 'В норме',
'last_visit': f"{randint(1, 31)}/{randint(1, 12)}/2020",
'last_data_update': f"{randint(1, 31)}/{randint(1, 12)}/2021",
'tmt_online': choice(['Online', 'Offline']),
'link': '/dashboard/diagnosis/index2' if pdf['Rhythms'] != 'Sinus rhythm' else '/dashboard/diagnosis',
},
print('{', end='')
for key, value in patient_data[0].items():
if key == 'avatar':
print(f"{key}: {value}", end=', ')
else:
print(f"{key}: '{value}'", end=', ')
print('},')
result['data'].append(patient_data)
except Exception as e:
print('Error (getting ecg data): {}'.format(e))
print(result)
result = json.dumps(result)
return result
def test_gqrs(self):
record = wfdb.rdrecord('sample-data/100', channels=[0],
sampfrom=9998, sampto=19998, physical=False)
expected_peaks = [271, 580, 884, 1181, 1469, 1770, 2055, 2339, 2634,
2939, 3255, 3551, 3831, 4120, 4412, 4700, 5000, 5299,
5596, 5889, 6172, 6454, 6744, 7047, 7347, 7646, 7936,
8216, 8503, 8785, 9070, 9377, 9682]
peaks = processing.gqrs_detect(d_sig=record.d_signal[:,0],
fs=record.fs,
adc_gain=record.adc_gain[0],
adc_zero=record.adc_zero[0],
threshold=1.0)
assert np.array_equal(peaks, expected_peaks)
def test_gqrs(self):
record = wfdb.rdrecord('sample-data/100', channels=[0],
sampfrom=9998, sampto=19998, physical=False)
expected_peaks = [271, 580, 884, 1181, 1469, 1770, 2055, 2339, 2634,
2939, 3255, 3551, 3831, 4120, 4412, 4700, 5000, 5299,
5596, 5889, 6172, 6454, 6744, 7047, 7347, 7646, 7936,
8216, 8503, 8785, 9070, 9377, 9682]
peaks = processing.gqrs_detect(d_sig=record.d_signal[:,0],
fs=record.fs,
adc_gain=record.adc_gain[0],
adc_zero=record.adc_zero[0],
threshold=1.0)
assert np.array_equal(peaks, expected_peaks)
def gqrs_algorithm(filename,
sampfrom=None,
sampto=None,
channel=0,
r_peak_inds=None,
fig=None,
pic_index=1,
pic_size=1,
skip_flag=False):
sig, fields = wfdb.rdsamp(filename,
channels=[channel],
sampfrom=sampfrom,
sampto=sampto)
qrs_inds = processing.gqrs_detect(sig=sig[:, 0], fs=fields['fs'])
# 标记位置减去采样点开始位置得到相对位置
r_peak_inds -= sampfrom
if len(qrs_inds) < 1:
qrs_inds = np.array([-100])
if skip_flag:
return draw_graph(r_peak_inds,
sig,
fields,
'GQRS',
qrs_inds,
fig=fig,
pic_index=pic_index,
pic_size=pic_size,
skip_flag=skip_flag)
summary, fig = draw_graph(r_peak_inds,
sig,
fields,
'GQRS',
qrs_inds,
fig=fig,
pic_index=pic_index,
pic_size=pic_size)
return summary, fig
def rpeak_gqrs(ecg, fs):
rr_min = 0.25
rr_max = 1.5
res = processing.gqrs_detect(sig=ecg, fs=fs, RRmin=rr_min, RRmax=rr_max)
return res
def ExtractFeatures(dataset, size, minThreshold):
# Features
qAmplitudes = []
rAmplitudes = []
qrsDurations = []
rrIntervals = []
heartRatesDuringHeartBeat = []
minSize = sys.maxsize
signals = dataset['signal']
labels = dataset['label']
for i, sig in enumerate(signals[:size]):
print(str(i) + "/" + str(len(signals)) + " ...")
qrsInds = processing.gqrs_detect(sig=sig.astype('float64'), fs=300)
heartRates = processing.compute_hr(sig_len=sig.shape[0],
fs=300,
qrs_inds=sorted(qrsInds))
rPoints, sPoints, qPoints = QRS_util.ECG_QRS_detect(
sig, 300, True, False)
lenHb = sys.maxsize
lenSigQrs = sys.maxsize
lenHr = sys.maxsize
lenQ = sys.maxsize
lenR = sys.maxsize
if (len(heartRates) > 0 and len(rPoints) > 0):
# Adding features for each signal.
heartRatesDuringHeartBeat.append(np.array(heartRates[rPoints]))
lenHr = len(heartRates[rPoints])
if (len(qPoints) > 0):
qAmplitudes.append(np.array(sig[qPoints]))
lenQ = len(sig[qPoints])
if (len(rPoints) > 0):
rAmplitudes.append(np.array(sig[rPoints]))
lenR = len(sig[rPoints])
if (len(qPoints) > 0 and len(rPoints) > 0):
sigQrsDuration = [
sPoints[i] - qPoints[i] for i in range(len(qPoints) - 1)
]
qrsDurations.append(np.array(sigQrsDuration))
lenSigQrs = len(sigQrsDuration)
hbIntervals = [
rPoints[i + 1] - rPoints[i] for i in range(len(rPoints) - 1)
]
rrIntervals.append(np.array(hbIntervals))
lenHb = len(hbIntervals)
minIter = min(lenHb, lenQ, lenR, lenSigQrs, lenHr)
if minIter < minThreshold:
# Rollback
del heartRatesDuringHeartBeat[-1]
del qAmplitudes[-1]
del rAmplitudes[-1]
del qrsDurations[-1]
del rrIntervals[-1]
elif minIter < minSize:
minSize = minIter
qAmplitudes = np.array(
[np.array(x[:minSize], dtype='int64') for x in qAmplitudes])
rAmplitudes = np.array(
[np.array(x[:minSize], dtype='int64') for x in rAmplitudes])
rrIntervals = np.array(
[np.array(x[:minSize], dtype='int64') for x in rrIntervals])
qrsDurations = np.array(
[np.array(x[:minSize], dtype='int64') for x in qrsDurations])
heartRatesDuringHeartBeat = np.array([
np.array(x[:minSize], dtype='int64') for x in heartRatesDuringHeartBeat
])
qAmplitudes = np.stack(qAmplitudes, axis=0)
rAmplitudes = np.stack(rAmplitudes, axis=0)
heartRatesDuringHeartBeat = np.stack(heartRatesDuringHeartBeat, axis=0)
rrIntervals = np.stack(rrIntervals, axis=0)
qrsDurations = np.stack(qrsDurations, axis=0)
return qAmplitudes, rAmplitudes, heartRatesDuringHeartBeat, rrIntervals, qrsDurations, labels[:
size], minSize
from imutils.video import VideoStream
from imutils import face_utils
from threading import Thread
import numpy as np
import pyglet
import argparse
import imutils
import time
import dlib
import cv2
# 1. Read signal
record = wfdb.rdrecord('mit-bih-arrhythmia-database-1.0.0/108', channels=[0])
# 2. Detect QRS locations
qrs_locs = processing.gqrs_detect(record.p_signal[:, 0], fs=record.fs)
# 5. Length of signal
len = record.sig_len
# 6. Compute heart rate
heart_rate = processing.compute_hr(len, qrs_locs, fs=record.fs)
# 7. Remove Nan values from array
heart_rate = heart_rate[~np.isnan(heart_rate)]
# 8. Calculate RR from heart rate
# RR-interval is shorter and wider for awake subject than drowsy subject.
RR = 60 / heart_rate
# From hrv analysis get frequency features
def gqrs_single(fields, record, save_path, sig, save=True):
r_peaks = processing.gqrs_detect(sig[:, 0], fs=fields['fs'])
if save:
save_prediction(r_peaks, record, save_path)
else:
return r_peaks
import wfdb
from wfdb import processing
import numpy as np
DATA_PATH = "/data/"
SAVE_PATH = "/pred/"
with open(DATA_PATH + "RECORDS", 'r') as f:
records = f.readlines()
records = list(map(lambda r: r.strip("\n"), records))
for record in records:
sig, fields = wfdb.rdsamp(DATA_PATH + record, channels=[0])
res = processing.gqrs_detect(sig[:, 0], fs=fields['fs'])
if len(res) > 0:
wfdb.wrann(record,
'atr',
res,
write_dir=SAVE_PATH,
symbol=(['N'] * len(res)))
def detect(self, records):
return [
processing.gqrs_detect(sig=record.p_signal.T[0], fs=record.fs)
for record in records
]
ax_left.set_title(title)
ax_left.set_xlabel('Time (ms)')
ax_left.set_ylabel('ECG (mV)', color='#3979f0')
ax_right.set_ylabel('Heart rate (bpm)', color='m')
#设置颜色使得和线条颜色一致
ax_left.tick_params('y', colors='#3979f0')
ax_right.tick_params('y', colors='m')
if saveto is not None:
plt.savefig(saveto, dpi=600)
plt.show()
#加载ECG信号
record=wfdb.rdrecord('./2') #.hea .dat文件名称
#help(wfdb.rdrecord)
#使用gqrs算法定位qrs波位置
qrs_inds=processing.gqrs_detect(sig=record.p_signal[:, 0], fs=record.fs) #未矫正位置
#画出结果
#peaks_hr(sig=record.p_signal, peak_inds=qrs_inds, fs=record.fs, title='GQRS peak detection on record 100')
#修正峰值，将其设置为局部最大值
min_bpm=20
max_bpm=230
#使用可能最大的bpm作为搜索半径
search_radius=int(record.fs*60/max_bpm)
corrected_peak_inds=processing.correct_peaks(record.p_signal[:, 0], peak_inds=qrs_inds, search_radius=search_radius, smooth_window_size=150)
#输出矫正后的QRS波峰位置
print('Corrected gqrs detected peak indices:', sorted(corrected_peak_inds))
# Feature 1: 计算R波波峰
def dataGeneration(data_path, csv_path, record_path):
# initialize dataset
dataset = pd.DataFrame(columns=['label', 'record'])
if record_path == None:
# a loop for each patient
detail_path = data_path + '/'
record_files = [
i.split('.')[0] for i in os.listdir(detail_path)
if (not i.startswith('.') and i.endswith('.hea'))
]
Bar.check_tty = False
bar = Bar('Processing',
max=len(record_files),
fill='#',
suffix='%(percent)d%%')
# a loop for each record
for record_name in record_files:
# load record
signal, info = wfdb.rdsamp(detail_path + record_name)
fs = 200
signal = processing.resample_sig(signal[:, 0], info['fs'], fs)[0]
# set some parameters
window_size_half = int(fs * 0.125 / 2)
max_bpm = 230
# detect QRS peaks
qrs_inds = processing.gqrs_detect(signal, fs=fs)
search_radius = int(fs * 60 / max_bpm)
corrected_qrs_inds = processing.correct_peaks(
signal,
peak_inds=qrs_inds,
search_radius=search_radius,
smooth_window_size=150)
average_qrs = 0
count = 0
for i in range(1, len(corrected_qrs_inds) - 1):
start_ind = corrected_qrs_inds[i] - window_size_half
end_ind = corrected_qrs_inds[i] + window_size_half + 1
if start_ind < corrected_qrs_inds[
i - 1] or end_ind > corrected_qrs_inds[i + 1]:
continue
average_qrs = average_qrs + signal[start_ind:end_ind]
count = count + 1
# remove outliers
if count < 8:
print('\noutlier detected, discard ' + record_name)
continue
average_qrs = average_qrs / count
corrcoefs = []
for i in range(1, len(corrected_qrs_inds) - 1):
start_ind = corrected_qrs_inds[i] - window_size_half
end_ind = corrected_qrs_inds[i] + window_size_half + 1
if start_ind < corrected_qrs_inds[
i - 1] or end_ind > corrected_qrs_inds[i + 1]:
corrcoefs.append(-100)
continue
corrcoef = pearsonr(signal[start_ind:end_ind], average_qrs)[0]
corrcoefs.append(corrcoef)
max_corr = list(map(corrcoefs.index, heapq.nlargest(8, corrcoefs)))
index_corr = random.sample(
list(itertools.permutations(max_corr, 8)), 100)
for index in index_corr:
# a temp dataframe to store one record
record_temp = pd.DataFrame()
signal_temp = []
for i in index:
start_ind = corrected_qrs_inds[i + 1] - window_size_half
end_ind = corrected_qrs_inds[i + 1] + window_size_half + 1
sig = processing.normalize_bound(signal[start_ind:end_ind],
-1, 1)
signal_temp = np.concatenate((signal_temp, sig))
record_temp = record_temp.append(pd.DataFrame(
signal_temp.reshape(-1, signal_temp.shape[0])),
ignore_index=True,
sort=False)
record_temp['label'] = record_name
record_temp['record'] = record_name
# add it to final dataset
dataset = dataset.append(record_temp,
ignore_index=True,
sort=False)
bar.next()
bar.finish()
else:
patient_folders = [
i for i in os.listdir(data_path)
if (not i.startswith('.') and i.startswith(record_path))
]
Bar.check_tty = False
bar = Bar('Processing',
max=len(patient_folders),
fill='#',
suffix='%(percent)d%%')
# a loop for each patient
for patient_name in patient_folders:
detail_path = data_path + patient_name + '/'
record_files = [
i.split('.')[0] for i in os.listdir(detail_path)
if i.endswith('.hea')
]
# a loop for each record
for record_name in record_files:
# load record
signal, info = wfdb.rdsamp(detail_path + record_name)
fs = 200
signal = processing.resample_sig(signal[:, 0], info['fs'],
fs)[0]
# set some parameters
window_size_half = int(fs * 0.125 / 2)
max_bpm = 230
# detect QRS peaks
qrs_inds = processing.gqrs_detect(signal, fs=fs)
search_radius = int(fs * 60 / max_bpm)
corrected_qrs_inds = processing.correct_peaks(
signal,
peak_inds=qrs_inds,
search_radius=search_radius,
smooth_window_size=150)
average_qrs = 0
count = 0
for i in range(1, len(corrected_qrs_inds) - 1):
start_ind = corrected_qrs_inds[i] - window_size_half
end_ind = corrected_qrs_inds[i] + window_size_half + 1
if start_ind < corrected_qrs_inds[
i - 1] or end_ind > corrected_qrs_inds[i + 1]:
continue
average_qrs = average_qrs + signal[start_ind:end_ind]
count = count + 1
# remove outliers
if count < 8:
print('\noutlier detected, discard ' + record_name +
' of ' + patient_name)
continue
average_qrs = average_qrs / count
corrcoefs = []
for i in range(1, len(corrected_qrs_inds) - 1):
start_ind = corrected_qrs_inds[i] - window_size_half
end_ind = corrected_qrs_inds[i] + window_size_half + 1
if start_ind < corrected_qrs_inds[
i - 1] or end_ind > corrected_qrs_inds[i + 1]:
corrcoefs.append(-100)
continue
corrcoef = pearsonr(signal[start_ind:end_ind],
average_qrs)[0]
corrcoefs.append(corrcoef)
max_corr = list(
map(corrcoefs.index, heapq.nlargest(8, corrcoefs)))
index_corr = random.sample(
list(itertools.permutations(max_corr, 8)), 100)
for index in index_corr:
# a temp dataframe to store one record
record_temp = pd.DataFrame()
signal_temp = []
for i in index:
start_ind = corrected_qrs_inds[i +
1] - window_size_half
end_ind = corrected_qrs_inds[i +
1] + window_size_half + 1
sig = processing.normalize_bound(
signal[start_ind:end_ind], -1, 1)
signal_temp = np.concatenate((signal_temp, sig))
record_temp = record_temp.append(pd.DataFrame(
signal_temp.reshape(-1, signal_temp.shape[0])),
ignore_index=True,
sort=False)
record_temp['label'] = patient_name
record_temp['record'] = record_name
# add it to final dataset
dataset = dataset.append(record_temp,
ignore_index=True,
sort=False)
bar.next()
bar.finish()
# save for further use
dataset.to_csv(csv_path, index=False)
print('processing completed')
