def make_fig2(save=False):
data_params = {}
data_params[
'path'] = '/research/vibs/Pedro/datasets/motorimagery/BCI-competitions/BCI-IV/2a/'
data_params['session'] = 1
data_params['task'] = 1
data_params['tparams'] = [1.25, 3.75]
data_params['fparams'] = [8.0, 35.0]
data_params['subject'] = 5
X, yworst = get_data(data_params)
X = X[(yworst == 1) | (yworst == 2)]
yworst = yworst[(yworst == 1) | (yworst == 2)]
covs = Covariances().fit_transform(X)
uworst, lworst = get_diffusionEmbedding(points=covs,
distance=distance_riemann)
data_params['subject'] = 1
X, ybest = get_data(data_params)
X = X[(ybest == 1) | (ybest == 2)]
ybest = ybest[(ybest == 1) | (ybest == 2)]
covs = Covariances().fit_transform(X)
ubest, lbest = get_diffusionEmbedding(points=covs,
distance=distance_riemann)
fig = plt.figure(figsize=(10.5, 5))
plt.subplots_adjust(wspace=0.020, hspace=0.025)
plt.subplot(1, 2, 1)
colorst = [['b', 'r'][int(t)] for t in (yworst - 2)]
plt.scatter(uworst[:, 1], uworst[:, 2], color=colorst, s=44)
plt.xlabel(r'$\phi_1$', fontsize=26)
plt.ylabel(r'$\phi_2$', fontsize=26)
plt.xticks([])
plt.yticks([])
ttl = plt.title('Worst subject', fontsize=20)
ttl.set_position([.5, 1.025])
ax = plt.subplot(1, 2, 2)
colorst = [['b', 'r'][int(t)] for t in (ybest - 2)]
plt.scatter(ubest[:, 1], ubest[:, 2], color=colorst, s=44)
plt.xlabel(r'$\phi_1$', fontsize=26)
plt.ylabel(r'$\phi_2$', fontsize=26)
ax.yaxis.set_label_position("right")
plt.xticks([])
plt.yticks([])
ttl = plt.title('Best subject', fontsize=20)
ttl.set_position([.5, 1.025])
if save:
name = 'figure2'
savefigure(name)
return [uworst, lworst], [ubest, lbest]
y = []
for _ in range(nt / 2):
x.append(gen_sig(ns, condition=2))
y.append(1)
for _ in range(nt / 2):
x.append(gen_sig(ns, condition=4))
y.append(2)
x = np.stack(x, axis=0)
#%%
import sys
sys.path.append('../diffusion/')
from diffusionmap import get_diffusionEmbedding
u_spectrum, _ = get_diffusionEmbedding(x, distance_spectrum)
covs = Covariances().fit_transform(x)
u_covariance, _ = get_diffusionEmbedding(covs, distance_riemann)
#%%
plt.figure(figsize=(13, 6))
plt.subplot(1, 2, 1)
for i in range(nt):
plt.scatter(u_spectrum[i, 1], u_spectrum[i, 2], color=['b', 'r'][y[i] - 1])
plt.subplot(1, 2, 2)
for i in range(nt):
plt.scatter(u_covariance[i, 1],
u_covariance[i, 2],
color=['b', 'r'][y[i] - 1])
points = points[idx]
theta = theta[idx]
return points, theta
def get_color(x):
# note that x goes from 0 to 1
colormap = plt.cm.gist_rainbow
return colormap(x)
Npoints = 150
distance = distance_euc
points, theta = gen_data(Npoints)
fig = plt.figure(figsize=(16, 7))
ax = fig.add_subplot(121, projection='3d')
for i, point in enumerate(points):
thetan = theta[i] / max(theta)
ax.scatter(point[0], point[1], point[2], color=get_color(thetan))
ax = fig.add_subplot(122)
u, s = get_diffusionEmbedding(points, distance, alpha=1.0)
for i in range(Npoints):
thetan = theta[i] / max(theta)
ax.scatter(s[1] * u[i, 1], s[2] * u[i, 2], color=get_color(thetan))
ax.axis('equal')
fs = struct['freq']
fini = 4.0
fend = 40.0
b, a = signal.butter(5, [fini / (fs / 2), fend / (fs / 2)], btype='bandpass')
for xt in X:
f, pxx = welch(xt, fs=fs)
xt = signal.filtfilt(b, a, xt)
#%%
from pyriemann.estimation import Covariances
covs = Covariances(estimator='oas').fit_transform(X)
from diffusionmap import get_diffusionEmbedding
u, l = get_diffusionEmbedding(points=covs, distance=distance_riemann)
#%%
fig = plt.figure(figsize=(11, 9.6))
ax = fig.add_subplot(111, projection='3d')
green = Color("green")
colors = list(green.range_to(Color("red"), int(np.max(lat) + 1)))
for i in range(len(covs)):
ax.scatter(u[i, 1],
u[i, 2],
u[i, 3],
color=colors[int(lat[i])].get_rgb(),
s=44)
ax.set_xlim(-3, +3)
filepaths = glob.glob(path + '*.jpg')
X = []
for filepath in filepaths:
I = imread(filepath, flatten=True)
I = I[::4, ::4]
X.append(I)
X = np.stack(X)
#%%
def distance_image(IX, IY):
return np.linalg.norm(IX-IY)
from diffusionmap import get_diffusionEmbedding
u,l = get_diffusionEmbedding(points=X, distance=distance_image)
#%%
for ifig in range(len(filepaths)):
fig = plt.figure(figsize=(23.8,9.4), facecolor='white')
plt.subplot(1,2,1)
for ui in u:
plt.scatter(ui[1], ui[2], color='k', s=25)
plt.scatter(u[ifig,1], u[ifig,2], s=160, color='r')
plt.xlim(-2,2)
plt.ylim(-2,3)
plt.subplot(1,2,2)
def make_fig3(save=False):
data_params = {}
data_params[
'path'] = '/research/vibs/Pedro/datasets/motorimagery/BCI-competitions/BCI-IV/2a/'
data_params['task'] = 1
data_params['tparams'] = [1.25, 3.75]
data_params['fparams'] = [8.0, 35.0]
data_params['subject'] = 1
data_params['session'] = 1
X1, y1 = get_data(data_params)
X1 = X1[(y1 == 1) | (y1 == 2)]
y1 = y1[(y1 == 1) | (y1 == 2)]
data_params['session'] = 2
X2, y2 = get_data(data_params)
X2 = X2[(y2 == 1) | (y2 == 2)]
y2 = y2[(y2 == 1) | (y2 == 2)]
X2 = np.delete(X2, (37), axis=0) # delete bad trial
y2 = np.delete(y2, (37), axis=0) # delete bad trial
y = np.concatenate((y1, y2), axis=0)
covs = Covariances().fit_transform(X1)
u1, l1 = get_diffusionEmbedding(points=covs, distance=distance_riemann)
covs = Covariances().fit_transform(X2)
u2, l2 = get_diffusionEmbedding(points=covs, distance=distance_riemann)
covs = Covariances().fit_transform(np.concatenate((X1, X2), axis=0))
u, l = get_diffusionEmbedding(points=covs, distance=distance_riemann)
fig1 = plt.figure(figsize=(15.5, 5))
plt.subplots_adjust(wspace=0.020, hspace=0.025)
ax1 = plt.subplot(1, 3, 1)
colorst = [['b', 'r'][int(t)] for t in (y1 - 2)]
for i, ui in enumerate(u1):
plt.scatter(ui[1], ui[2], color=colorst[i], s=44)
ax1.set_xticks([])
ax1.set_yticks([])
plt.xlabel(r'$\phi_1$', fontsize=26)
plt.ylabel(r'$\phi_2$', fontsize=26)
ttl = plt.title('Session 1', fontsize=24)
ttl.set_position([.5, 1.025])
ax2 = plt.subplot(1, 3, 2)
colorst = [['b', 'r'][int(t)] for t in (y2 - 2)]
for i, ui in enumerate(u2):
plt.scatter(ui[1], ui[2], color=colorst[i], s=44)
ax2.set_xticks([])
ax2.set_yticks([])
plt.xlabel(r'$\phi_1$', fontsize=26)
ttl = plt.title('Session 2', fontsize=24)
ttl.set_position([.5, 1.025])
ax3 = plt.subplot(1, 3, 3)
colorst = [['b', 'r'][int(t)] for t in (y - 2)]
markerst = ['o' for _ in range(len(y1))] + ['*' for _ in range(len(y2))]
for i, ui in enumerate(u):
plt.scatter(ui[1], ui[2], color=colorst[i], s=44, marker=markerst[i])
ax3.set_xticks([])
ax3.set_yticks([])
plt.xlabel(r'$\phi_1$', fontsize=26)
plt.ylabel(r'$\phi_2$', fontsize=26)
ax3.yaxis.set_label_position("right")
ttl = plt.title('Both sessions', fontsize=24)
ttl.set_position([.5, 1.025])
if save:
name = 'figure3a'
savefigure(name)
fig2 = plt.figure(figsize=(9, 8))
ax = fig2.add_subplot(111, projection='3d')
colorst = [['b', 'r'][int(t)] for t in (y - 2)]
markerst = ['o' for _ in range(len(y1))] + ['*' for _ in range(len(y2))]
for i, ui in enumerate(u):
ax.scatter(ui[1],
ui[2],
ui[3],
color=colorst[i],
s=44,
marker=markerst[i])
ax.set_xlabel(r'$\phi_1$', fontsize=26)
ax.set_ylabel(r'$\phi_2$', fontsize=26)
ax.set_zlabel(r'$\phi_3$', fontsize=26)
ax.set_xticks([])
ax.set_yticks([])
ax.set_zticks([])
ttl = plt.title('Both sessions', fontsize=24)
ttl.set_position([0.50, 0.975])
if save:
name = 'figure3b'
savefigure(name)
# plt.ylim(0,20)
# ax = plt.gca()
#% Sliding window for covariances and diffusion embedding
L = 250
nt,nc,ns = X.shape
covm = []
for w in tqdm(gen_windows(L, ns, step=32)):
xw = X[:,:,w]
covs = Covariances().fit_transform(xw)
covm.append(mean_riemann(covs))
print 'getting the diffusion embedding'
covm = np.stack(covm)
u,l = get_diffusionEmbedding(covm, distance_riemann, alpha=1.0, tdiff=0)
filepath = 'trajectory_real_bci-iv_subject' + str(subject) + '.pkl'
embedding = [u,l]
joblib.dump(embedding, filepath)
#%% Visualization
state0 = np.array([+1, +1, +1])
state1 = np.array([-1, -1, -1])
sigma = 0.50
state0_pertub = np.array(
[state0 + sigma * np.random.randn(3) for _ in range(Npoints / 2)])
state1_pertub = np.array(
[state1 + sigma * np.random.randn(3) for _ in range(Npoints / 2)])
points = np.concatenate((state0_pertub, state1_pertub), axis=0)
theta = np.zeros(Npoints)
theta[Npoints / 2:] = 1
return points, theta
Npoints = 200
distance = distance_euc
points, theta = gen_data(Npoints)
u, s = get_diffusionEmbedding(points, distance)
fig = plt.figure(figsize=(18, 7))
colorst = [['b', 'r'][int(t)] for t in theta]
ax = fig.add_subplot(121, projection='3d')
ax.scatter(points[:, 0], points[:, 1], points[:, 2], color=colorst)
ax = fig.add_subplot(122)
ax.scatter(range(Npoints), u[:, 1], color=colorst)
