def __data_generation(self, list_IDs_temp):
'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)
# Initialization
X = np.empty((self.batch_size, *self.dim, self.n_channels), dtype = float)
y = np.empty((self.batch_size), dtype = int)
# Generate data
for i, ID in enumerate(list_IDs_temp):
data = extend_ts(self.h5file[ID]['ecgdata'][:, 0], self.sequence_length)
data = np.reshape(data, (1, len(data)))
if self.augment:
# dropout bursts
data = zero_filter(data, threshold = 2, depth = 10)
# random resampling
data = random_resample(data)
# Generate spectrogram
data_spectrogram = spectrogram(data, nperseg = self.nperseg, noverlap = self.noverlap)[2]
# Normalize
data_transformed = norm_float(data_spectrogram, self.data_mean, self.data_std)
X[i,] = np.expand_dims(data_transformed, axis = 3)
# Assuming that the dataset names are unique (only 1 per label)
y[i] = self.labels[ID]
return X, keras.utils.to_categorical(y, num_classes=self.n_classes)
def __data_generation(self, list_IDs_temp):
'Generates data containing batch_size samples' # X : (n_samples, *dim, n_channels)
# Initialization
X = np.empty((self.batch_size, *self.dim, self.n_channels), dtype=float)
y = np.empty((self.batch_size), dtype=int)
# Generate data
for i, ID in enumerate(list_IDs_temp):
data = extend_ts(self.h5file[ID]['ecgdata'][:, 0],
self.sequence_length)
data = np.reshape(data, (1, len(data)))
if self.augment:
# dropout bursts
data = zero_filter(data, threshold=2, depth=10)
# random resampling
data = random_resample(data)
# Generate spectrogram
data_spectrogram = spectrogram(data,
nperseg=self.nperseg,
noverlap=self.noverlap)[2]
# Normalize spectrogram
#data_transformed = norm_float(data_spectrogram, self.data_mean, self.data_std)
data_norm = (data_spectrogram -
np.mean(data_spectrogram)) / np.std(data_spectrogram)
X[i, ] = np.expand_dims(data_norm, axis=3)
# Assuming that the dataset names are unique (only 1 per label)
y[i] = self.labels[ID]
return X, keras.utils.to_categorical(y, num_classes=self.n_classes)
f.set(ylabel='Counts', ylim=[0, 200], yticks=np.arange(0, 250, 50))
plt.title('Distribution of sequence lengths', fontsize=15)
plt.show()
fig.savefig('physionet_sequenceLenHist.png', bbox_inches='tight', dpi=150)
# Based on this, we can set some parameters that we will use in the future
fs = sampling_rates[0] # universal sampling rate
sequence_length = sequence_length_max # will use the maximum sequence length
from physionet_processing import extend_ts
ts = h5file[
dataset_list[15]]['ecgdata'][:,
0] # Fetch one time series from the hdf5 file
#ts = h5file[list(h5file.keys())[20]]['ecgdata']
ts_extended = extend_ts(
ts, length=sequence_length_max) # Extend it to the maximum length
time = np.arange(0, len(ts_extended)) / fs
# Plot the the extended time series
fig, ax1 = plt.subplots(figsize=(15, 3))
ax1.plot(time, ts_extended, 'b')
ax1.set(xlabel='Time [s]',
xlim=[0, time[-1]],
xticks=np.arange(0, time[-1] + 5, 10))
ax1.set(ylabel='Potential [mV]')
plt.title('Example ECG sequence with zero padding', fontsize=15)
fig.savefig('physionet_ECG_padding.png', bbox_inches='tight', dpi=150)
plt.show()
from scipy import signal
f1, PSD = signal.periodogram(ts_extended, fs, 'flattop', scaling='density')