def polyfit_test():
image = load_images(0, 1, False, 5)
spr = spread(image)[0]
p = polyfit(image, 10)[0]
plt.scatter(range(0, len(spr)), spr)
plt.scatter(range(0, len(spr)), np.polyval(p, range(0, len(spr))))
plt.show()
def test_max_sobel():
images = load_images(0, 50, False, 2)
sobels = sobel(images, lambda x: x)
ages = load_ages()
for i in range(0, len(images)):
print sobels[i].max()
print ages[i]
def plot_spread_sobel():
images = load_images(0, 10, False, 2)
spreads = feature_on_cerebrum(images, feature_function=sobel)
ages = load_ages()
_, ax = plt.subplots(2, 5, sharex=True, sharey=True)
ax = ax.flatten()
for i in range(0, len(images)):
ax[i].scatter(range(0, len(spreads[i])), spreads[i], s=1, marker=",")
ax[i].set_title("healthy" if ages[i] == 1 else "ill")
plt.show()
def plot_derivatives():
images = load_images(0, 10, False, 5)
derivs = derivatives(images)
ages = load_ages()
_, ax = plt.subplots(2, 5, sharex=True, sharey=True)
ax = ax.flatten()
for i in range(0, len(images)):
ax[i].scatter(range(0, len(derivs[i])), derivs[i], s=1, marker=",")
ax[i].set_title("healthy" if ages[i] == 1 else "ill")
plt.show()
def smooth_sample_test():
image = load_images(0, 1, False, 5)
spr = spread(image)[0]
smoothed0 = smooth(spr, 51, "hanning")
smoothed1 = smooth(spr, 51, "hanning")
smoothed2 = smooth(spr, 25, "hanning")
plt.scatter(range(0, len(spr)), spr) # blue
plt.scatter(range(0, len(spr)), smoothed1) # black
# plt.scatter(range(0, len(spr)), smoothed2)
plt.show()
def plot_median_sobel():
images = load_images(0, 50, False, 2)
mean = sobel(images, lambda x: [[np.mean(xi)] for xi in x])
vars = sobel(images, variance_feature)
ages = load_ages()
for i in range(0, len(images)):
plt.scatter(median[i][0],
variance[i][0],
s=1,
marker=",",
color="blue" if ages[i] == 1 else "red")
plt.show()
def plot_sobel():
images = load_images(0, 1, False, 2)
image = images[0]
print image.shape
sx = sp.ndimage.filters.sobel(image, axis=0)
print sx.max()
sy = sp.ndimage.filters.sobel(image, axis=1)
print sy.max()
sz = sp.ndimage.filters.sobel(image, axis=2)
print sz.max()
sob = np.sqrt(sx * sx + sy * sy + sz * sz)
sob = sob[:, :, :, 0] # unpack
print sob.shape
sob = sob[40]
plt.imshow(sob, cmap="gray")
plt.show()
def find_local_maxima_test():
images = load_images(0, 5, False, 5)
spread_data = spread(images)
maxima = local_maxima_without_smoothing(images)
smoothed_maxima = local_maxima_3(images)
print maxima
print smoothed_maxima
_, ax = plt.subplots(1, 5, sharey=True)
ax = ax.flatten()
for i in range(5):
smoothed = smooth(spread_data[i], window_len=101)
ax[i].scatter(range(len(spread_data[i])),
spread_data[i],
s=1,
marker=",")
ax[i].scatter(range(len(spread_data[i])), smoothed)
ax[i].set_title(maxima[i])
plt.ylim([0, 40])
plt.show()
def plot_half_mask_test():
images = load_images(0, 1, False, 2)
border_mask = find_border_mask(images[0])
mask = border_mask - get_shaved_mask(border_mask, 12)
heigth = 45
mask = mask[:, :, -heigth:]
print np.unique(mask)
print np.sum(mask)
xs = []
ys = []
zs = []
for x in range(len(mask) / 2):
for y in range(len(mask[0])):
for z in range(len(mask[0][0])):
if mask[x][y][z][0] == 1:
xs.append(x)
ys.append(y)
zs.append(z)
xs = np.array(xs)
ys = np.array(ys)
zs = np.array(zs)
fig = plt.figure()
ax = fig.gca(projection='3d')
ax.set_aspect('equal')
ax.scatter(xs, ys, zs)
max_range = np.array(
[xs.max() - xs.min(),
ys.max() - ys.min(),
zs.max() - zs.min()]).max() / 2.0
mid_x = (xs.max() + xs.min()) * 0.5
mid_y = (ys.max() + ys.min()) * 0.5
mid_z = (zs.max() + zs.min()) * 0.5
ax.set_xlim(mid_x - max_range, mid_x + max_range)
ax.set_ylim(mid_y - max_range, mid_y + max_range)
ax.set_zlim(mid_z - max_range, mid_z + max_range)
plt.show()
def local_maxima_3_test():
image = load_images(0, 1, False, 5)
maxima = local_maxima_3(image)
print maxima
def mean_feature_test():
image = load_images(0, 1, False, 5)
means = mean_feature(image)
print means
def mask_similarity_test():
images = load_images(0, 10, False, 1)
for i in range(len(images) - 1):
sum_of_difference = np.sum(
find_border_mask(images[i]) - find_border_mask(images[i + 1]))
print sum_of_difference