def test_train_segments():
results = np.zeros((6, 12))
for i in range(6):
dat = io.loadmat('testset%02d.mat' % (i + 1))
eog_v = dat['eog_v_seg'][0]
eog_h = dat['eog_h_seg'][0]
fs = float(dat['fs'][0])
fir_len = 150
stop_limit = 500 / fs
wn = stop_limit / (fs / 2)
b = sig.firwin(fir_len, wn)
eog_v_f = sig.filtfilt(b, 1, eog_v)
eog_h_f = sig.filtfilt(b, 1, eog_h)
p = PyGERT()
p.train(eog_h_f, eog_v_f)
results[i, 0] = p.mu_fix
results[i, 1] = p.sigma_fix
results[i, 2] = p.prior_fix
results[i, 3] = p.mu_sac
results[i, 4] = p.sigma_sac
results[i, 5] = p.prior_sac
results[i, 6] = p.mu_bli
results[i, 7] = p.sigma_bli
results[i, 8] = p.prior_bli
results[i, 9] = p.mu_bs
results[i, 10] = p.sigma_bs
results[i, 11] = p.prior_bs
print(results)
np.savetxt('segment_test_python.csv', results, delimiter=',')
def test_train_segments():
results = np.zeros((6, 12))
for i in range(6):
dat = io.loadmat('testset%02d.mat' % (i + 1))
eog_v = dat['eog_v_seg'][0]
eog_h = dat['eog_h_seg'][0]
fs = float(dat['fs'][0])
fir_len = 150
stop_limit = 500 / fs
wn = stop_limit / (fs/2)
b = sig.firwin(fir_len, wn)
eog_v_f = sig.filtfilt(b, 1, eog_v)
eog_h_f = sig.filtfilt(b, 1, eog_h)
p = PyGERT()
p.train(eog_h_f, eog_v_f)
results[i, 0] = p.mu_fix
results[i, 1] = p.sigma_fix
results[i, 2] = p.prior_fix
results[i, 3] = p.mu_sac
results[i, 4] = p.sigma_sac
results[i, 5] = p.prior_sac
results[i, 6] = p.mu_bli
results[i, 7] = p.sigma_bli
results[i, 8] = p.prior_bli
results[i, 9] = p.mu_bs
results[i, 10] = p.sigma_bs
results[i, 11] = p.prior_bs
print(results)
np.savetxt('segment_test_python.csv', results, delimiter=',')
def test_train_fulldata():
eog_h_f, eog_v_f = prepare_data()
p = PyGERT()
p.train(eog_h_f, eog_v_f)
print(' \tmu\tsd\tpri')
print('sac:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_sac, p.sigma_sac, p.prior_sac))
print('bli:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_bli, p.sigma_bli, p.prior_bli))
print('fix:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_fix, p.sigma_fix, p.prior_fix))
print('bs :\t%0.2f\t%0.2f\t%0.2f' % (p.mu_bs, p.sigma_bs, p.prior_bs))
def test_train_fulldata():
eog_h_f, eog_v_f = prepare_data()
p = PyGERT()
p.train(eog_h_f, eog_v_f)
print(' \tmu\tsd\tpri')
print('sac:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_sac, p.sigma_sac, p.prior_sac))
print('bli:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_bli, p.sigma_bli, p.prior_bli))
print('fix:\t%0.2f\t%0.2f\t%0.2f' % (p.mu_fix, p.sigma_fix, p.prior_fix))
print('bs :\t%0.2f\t%0.2f\t%0.2f' % (p.mu_bs, p.sigma_bs, p.prior_bs))
def offline_test_package(train_file='train.mat', detect_file='online.mat'):
prints('Starting offline test')
pg = PyGERT(stream_name=None)
prints('Loading data')
data = io.loadmat(train_file)
eog_h = data['EOGh'][0]
eog_v = data['EOGh'][0]
prints('Filtering data')
eog_h = sig.filtfilt(pg._b1, 1, eog_h)
eog_v = sig.filtfilt(pg._b1, 1, eog_v)
prints('Training PyGERT-instance')
training_ok = pg._train(eog_h, eog_v)
if training_ok:
prints('Training finished ok.')
else:
prints('Training had errors, re-run.')
# Save training params to file as well?
print(' \tmu\tsd\tpri')
print('sac:\t%0.2f\t%0.2f\t%0.2f' %
(pg.mu_sac, pg.sigma_sac, pg.prior_sac))
print('bli:\t%0.2f\t%0.2f\t%0.2f' %
(pg.mu_bli, pg.sigma_bli, pg.prior_bli))
print('fix:\t%0.2f\t%0.2f\t%0.2f' %
(pg.mu_fix, pg.sigma_fix, pg.prior_fix))
print('bs :\t%0.2f\t%0.2f\t%0.2f' % (pg.mu_bs, pg.sigma_bs, pg.prior_bs))
training_results = [
pg.mu_sac, pg.sigma_sac, pg.prior_sac, pg.mu_bli, pg.sigma_bli,
pg.prior_bli, pg.mu_fix, pg.sigma_fix, pg.prior_fix, pg.mu_bs,
pg.sigma_bs, pg.prior_bs
]
training_results = np.asarray(training_results)
np.savetxt('pygert_training.csv', training_results, delimiter=',')
prints('Running detection')
data = io.loadmat(detect_file)
eog_h = data['EOGh'][0]
eog_v = data['EOGv'][0]
results = np.ndarray((0, 3))
for h, v in zip(eog_h, eog_v):
h1, v1, h2, v2 = pg._filter_sample(h, v)
p = pg._detect(h1, v1, h2, v2)
results = np.vstack((results, p))
np.savetxt('pygert_results.csv', results, delimiter=',')
prints('Detection finished ok.')
def offline_test_package(train_file='train.mat', detect_file='online.mat'):
prints('Starting offline test')
pg = PyGERT(stream_name=None)
prints('Loading data')
data = io.loadmat(train_file)
eog_h = data['EOGh'][0]
eog_v = data['EOGh'][0]
prints('Filtering data')
eog_h = sig.filtfilt(pg._b1, 1, eog_h)
eog_v = sig.filtfilt(pg._b1, 1, eog_v)
prints('Training PyGERT-instance')
training_ok = pg._train(eog_h, eog_v)
if training_ok:
prints('Training finished ok.')
else:
prints('Training had errors, re-run.')
# Save training params to file as well?
print(' \tmu\tsd\tpri')
print('sac:\t%0.2f\t%0.2f\t%0.2f' % (pg.mu_sac, pg.sigma_sac,
pg.prior_sac))
print('bli:\t%0.2f\t%0.2f\t%0.2f' % (pg.mu_bli, pg.sigma_bli,
pg.prior_bli))
print('fix:\t%0.2f\t%0.2f\t%0.2f' % (pg.mu_fix, pg.sigma_fix,
pg.prior_fix))
print('bs :\t%0.2f\t%0.2f\t%0.2f' % (pg.mu_bs, pg.sigma_bs,
pg.prior_bs))
training_results = [pg.mu_sac, pg.sigma_sac, pg.prior_sac, pg.mu_bli,
pg.sigma_bli, pg.prior_bli, pg.mu_fix, pg.sigma_fix,
pg.prior_fix, pg.mu_bs, pg.sigma_bs, pg.prior_bs]
training_results = np.asarray(training_results)
np.savetxt('pygert_training.csv', training_results, delimiter=',')
prints('Running detection')
data = io.loadmat(detect_file)
eog_h = data['EOGh'][0]
eog_v = data['EOGv'][0]
results = np.ndarray((0, 3))
for h, v in zip(eog_h, eog_v):
h1, v1, h2, v2 = pg._filter_sample(h, v)
p = pg._detect(h1, v1, h2, v2)
results = np.vstack((results, p))
np.savetxt('pygert_results.csv', results, delimiter=',')
prints('Detection finished ok.')
