mirror of
https://github.com/This-weiyuqing/ECG_by_Diffenert_CNN_Test.git
synced 2026-10-18 03:13:39 +08:00
112 lines
2.8 KiB
Python
112 lines
2.8 KiB
Python
import numpy as np
|
|
import matplotlib.pyplot as plt
|
|
from numpy import reshape
|
|
|
|
# arr =
|
|
PATH = "mit-bih-arrhythmia-database-1.0.0/"
|
|
HEADERFILE = "100.hea"
|
|
ATRFILE = '100.atr'
|
|
DATAFILE = "100.dat"
|
|
|
|
# Obtain the information of each part from the hea file and fill in
|
|
f = open(PATH + HEADERFILE, "r")
|
|
z = f.readline().split()
|
|
nosig, sfreq = int(z[1]), int(z[2]) # single number,samp number
|
|
SAMPLES2READ = 10 * sfreq # number of data points
|
|
|
|
dformat, gain, bitres, zerovalue, firstvalue = [], [], [], [], []
|
|
|
|
for i in range(nosig):
|
|
z = f.readline().split()
|
|
dformat.append(int(z[1]))
|
|
gain.append(int(z[2]))
|
|
bitres.append(int(z[3]))
|
|
zerovalue.append(int(z[4]))
|
|
firstvalue.append(int(z[5]))
|
|
f.close()
|
|
|
|
# Obtain the information of each part from the dat file and fill in
|
|
f=open(PATH+DATAFILE, "rb")
|
|
b = f.read()
|
|
f.close()
|
|
|
|
A_init = np.frombuffer(b, dtype=np.uint8)
|
|
A_shape0 = int(A_init.shape[0] / 3)
|
|
A = A_init.reshape(A_shape0, 3)[:SAMPLES2READ]
|
|
|
|
M = np.zeros((SAMPLES2READ, 2))
|
|
|
|
M2H = A[:, 1] >> 4
|
|
M1H = A[:, 1] & 15
|
|
|
|
PRL = (A[:, 1] & 8) * (2 ** 9)
|
|
PRR = A[:, 1] & 128 << 5
|
|
|
|
M1H = M1H * (2 ** 8)
|
|
M2H = M2H * (2 ** 8)
|
|
|
|
M[:, 0] = A[:, 0] + M1H - PRL
|
|
M[:, 1] = A[:, 2] + M2H - PRR
|
|
|
|
if ((M[1, :] != firstvalue).any()):
|
|
print("inconsistent in the first bit value")
|
|
|
|
if nosig == 2:
|
|
M[:, 0] = (M[:, 0] - zerovalue[0]) / gain[0]
|
|
M[:, 1] = (M[:, 1] - zerovalue[1]) / gain[1]
|
|
TIME = np.linspace(0, SAMPLES2READ - 1, SAMPLES2READ) / sfreq
|
|
elif nosig == 1:
|
|
M2 = []
|
|
M[:, 0] = M[:, 0] - zerovalue[0]
|
|
M[:, 1] = M[:, 1] - zerovalue[1]
|
|
for i in range(M.shape[0]):
|
|
M2.append(M[:, 0][i])
|
|
M2.append(M[:, 1][i])
|
|
M2.append(0)
|
|
del M2[0]
|
|
M2 = np.array(M2) / gain[0]
|
|
TIME = np.linspace(0, 2 * SAMPLES2READ - 1, 2 * SAMPLES2READ) / sfreq
|
|
else:
|
|
print("Sorting algorithm for more than 2 signals not programmed yet")
|
|
|
|
# atr deal
|
|
f = open(PATH + ATRFILE, "rb")
|
|
b = f.read()
|
|
f.close()
|
|
A_init = np.frombuffer(b, dtype=np.uint8)
|
|
A_shape0 = int(A_init.shape[0] / 2)
|
|
A = A_init.reshape(A_shape0, 2)
|
|
|
|
ANNOT, ATRTIME = [], []
|
|
i = 0
|
|
while i < A.shape[0]:
|
|
annoth = A[i, 1] >> 2
|
|
if annoth == 59:
|
|
ANNOT.append(A[i + 3, 1] >> 2)
|
|
ATRTIME.append(A[i + 2, 0] + A[i + 2, 1] * (2 ** 8) + A[i + 1, 0] * (2 ** 16) + A[i + 1, 1] * (2 ** 24))
|
|
i += 3
|
|
elif annoth == 60:
|
|
pass
|
|
elif annoth == 61:
|
|
pass
|
|
elif annoth == 62:
|
|
pass
|
|
elif annoth == 63:
|
|
hilfe = (A[i, 1] & 3) * (2 * 8) + A[i, 0]
|
|
hilfe = hilfe + hilfe % 2
|
|
i += int(hilfe / 2)
|
|
else:
|
|
ATRTIME.append((A[i, 1] & 3) * (2 * 8) + A[i, 0])
|
|
ANNOT.append((A[i,1]>>2))
|
|
i+=1
|
|
del ANNOT[len(ANNOT)-1]
|
|
del ATRTIME[len(ATRTIME)-1]
|
|
|
|
ATRTIME=np.array(ATRTIME)
|
|
ATRTIME=np.cumsum(ATRTIME)/sfreq
|
|
|
|
ind=np.where(ATRTIME<=TIME[-1][0])
|
|
ATRTIMED=ATRTIME[ind]
|
|
|
|
ANNOT=np.round(ANNOT)
|
|
ANNOTD=ANNOT[ind] |