def plot_weights(weights):
w_c = weights.sum(axis=1) # 50x1
weights = weights / w_c.reshape((weights.shape[0], 1))
IMG = tile_raster_images(
X=weights,
img_shape=(28, 28), tile_shape=(10, 5),
tile_spacing=(1, 1))
show_image(IMG)
X_train, X_test = load_cars()
model = build_model()
if not False:
model.summary()
model.fit({
'input': X_train,
'autoencoder_feedback': X_train
},
nb_epoch=100,
batch_size=64,
validation_split=0.2,
callbacks=[EarlyStopping(patience=12)])
model.save_weights('./cars.neuro', overwrite=True)
else:
model.load_weights('./cars.neuro')
l = model.predict({'input': X_test[:25, ...]})
representations = np.clip(l['autoencoder_feedback'], 0, 1)
_r = tile_raster_images(X=keras2rgb(representations),
img_shape=(32, 32, 3),
tile_shape=(5, 5),
tile_spacing=(1, 1))
_o = tile_raster_images(X=keras2rgb(X_test),
img_shape=(32, 32, 3),
tile_shape=(5, 5),
tile_spacing=(1, 1))
show_image([(_o, 'Source'), (_r, 'Representations')])
# -*- coding: utf-8 -*-
""" exposure.py
Basic functions in scikit-image and matplotlib
"""
from skimage.io import imread
from skimage import exposure
import numpy as np
import matplotlib.pyplot as plt
from helpers import show_image, save_image
# Load Lena
lena = imread("../lena.jpg")
# Red channel
red_lena = lena[:,:,0]
# Plot histogram and choose
fig, (ax0, ax1) = plt.subplots(1, 2, figsize=(13.5,6))
ax0.hist(red_lena.ravel(), bins=100)
ax1.imshow(red_lena, cmap="gray", clim=(60.0, 240.0))
plt.show()
# Automatic exposure equalization
equal_lena = exposure.equalize_hist(red_lena, nbins=256)
show_image(equal_lena, colormap="gray")
#model.compile('rmsprop', loss='mean_squared_error')
return model
if __name__ == '__main__':
X_train, X_test = load_cars(python_version=sys.version_info.major)
model = build_model_functional()
if not False:
model.summary()
model.fit(X_train, X_train, nb_epoch=100, batch_size=64,
validation_split=0.2,
callbacks=[EarlyStopping(patience=12)])
model.save_weights('./cars.neuro', overwrite=True)
else:
model.load_weights('./cars.neuro')
l = model.predict(X_test[:25, ...])
representations = np.clip(l, 0, 1)
_r = tile_raster_images(
X=keras2rgb(representations),
img_shape=(32, 32, 3), tile_shape=(5, 5),
tile_spacing=(1, 1))
_o = tile_raster_images(
X=keras2rgb(X_test),
img_shape=(32, 32, 3), tile_shape=(5, 5),
tile_spacing=(1, 1))
show_image([(_o, 'Source'), (_r, 'Representations')])
import torch
import torchvision
from Classifier import Classifier
from classes import classes
from helpers import show_image, compute_accuracy, get_test_set_and_loader
if __name__ == '__main__':
torch.multiprocessing.freeze_support()
_, test_loader = get_test_set_and_loader()
dataiter = iter(test_loader)
images, labels = dataiter.next()
show_image(torchvision.utils.make_grid(images))
print('GroundTruth: ',
' '.join('%5s' % classes[labels[j]] for j in range(4)))
classifier = Classifier()
classifier.load_state_dict(torch.load('./model_storage/cifar_net.pth'))
outputs = classifier(images)
_, predicted = torch.max(outputs, 1)
print('Predicted: ',
' '.join('%5s' % classes[predicted[j]] for j in range(4)))
compute_accuracy(test_loader, classifier)
""" filters.py
Basic functions in scikit-image and matplotlib
"""
from skimage.io import imread
from skimage.filters import threshold_otsu
from skimage import filters
import numpy as np
from helpers import show_image, save_image
# Load Lena
lena = imread("../lena.jpg")
# Save the red channel
red_lena = lena[:, :, 0]
lena_1 = np.copy(red_lena)
# set values < mean intensity to 0
mean_intensity = np.mean(red_lena)
lena_1[red_lena < mean_intensity] = 250
show_image(lena_1, colormap="gray")
# Thresholding with Otsu's method
thresh = threshold_otsu(red_lena)
lena_2 = red_lena > thresh
show_image(lena_2, colormap="gray")
ratio = image.shape[0] / float(resized.shape[0])
b, g, r = cv2.split(resized)
# show_image("blue", b)
# show_image("green", g)
# show_image("red", r)
avg_red = np.average(r)
avg_green = np.average(g)
avg_blue = np.average(b)
thresh_red = cv2.threshold(r, avg_red + 10, 255, cv2.THRESH_BINARY)[1]
thresh_green = cv2.threshold(g, avg_green + 10, 255, cv2.THRESH_BINARY)[1]
thresh_blue = cv2.threshold(b, avg_blue + 10, 255, cv2.THRESH_BINARY)[1]
show_image("blue", thresh_blue)
show_image("green", thresh_green)
show_image("red", thresh_red)
for item in [thresh_green, thresh_blue, thresh_red]:
contours = cv2.findContours(item.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
contours = contours[0] if imutils.is_cv2() else contours[1]
sd = ShapeDetector()
# loop over the contours
for contour in contours:
print("contour")
# compute the center of the contour, then detect the name of the
# shape using only the contour
from skimage.morphology import disk
from skimage.filters.rank import mean
from skimage.color import rgb2gray
from scipy import fftpack
from helpers import show_image, save_image
# Load Lena and Lena_with_glasses
lena = imread("../lena.jpg")
lena_glass = imread("../lena_glass.jpg")
# ====== Point operators =======
# ------ Pixelweise subtraction ------
glass = lena - lena_glass
show_image(glass)
# ====== Local operators =======
# ------ Mean filter ------
mean_lena = mean(rgb2gray(lena), disk(5))
show_image(mean_lena, colormap="gray")
# ====== Global operators ======
# ------ Fourier transformation ------
F1 = fftpack.fft2(rgb2gray(lena))
F2 = fftpack.fftshift(F1)
F2_copy = np.copy(F2)
# Calculate a 2D power spectrum
psd2D = np.abs(F2)
# gray = cv2.cvtColor(resized, cv2.COLOR_BGR2GRAY)
# blurred = cv2.GaussianBlur(resized, (3, 3), 0)
cannied = imutils.auto_canny(v)
dilation = cv2.dilate(cannied,
cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5)),
iterations=5)
closing = cv2.morphologyEx(dilation, cv2.MORPH_CLOSE,
cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5)))
erosion = cv2.erode(closing,
cv2.getStructuringElement(cv2.MORPH_RECT, (5, 5)),
iterations=3)
# show_image("Input image", display_image)
# show_image("Grayscale image", gray)
# show_image("Blurred image", blurred)
show_image("Cannied image", cannied)
show_image("Closed image", erosion)
# thresh = cv2.threshold(blurred, avg_gray_color, 255, cv2.THRESH_BINARY)[1]
# find contours in the thresholded image and initialize the shape detector
contours = cv2.findContours(erosion.copy(), cv2.RETR_EXTERNAL,
cv2.CHAIN_APPROX_SIMPLE)
contours = contours[0] if imutils.is_cv2() else contours[1]
# Find biggest contour
biggest = None
biggest_size = 0
for contour in contours:
rect = cv2.minAreaRect(contour)
""" filters.py
Basic functions in scikit-image and matplotlib
"""
from skimage.io import imread
from skimage.filters import threshold_otsu
from skimage import filters
import numpy as np
from helpers import show_image, save_image
# Load Lena
lena = imread("../lena.jpg")
# Save the red channel
red_lena = lena[:,:,0]
lena_1 = np.copy(red_lena)
# set values < mean intensity to 0
mean_intensity = np.mean(red_lena)
lena_1[red_lena<mean_intensity] = 250
show_image(lena_1, colormap="gray")
# Thresholding with Otsu's method
thresh = threshold_otsu(red_lena)
lena_2 = red_lena > thresh
show_image(lena_2, colormap="gray")
#!/usr/bin/env python2
# -*- coding: utf-8 -*-
"""
Created on Wed Apr 18 14:03:21 2018
@author: wisse
"""
from helpers import show_image
import numpy as np
data = np.loadtxt('speedUp.csv', delimiter=',', dtype=object)
for obs in data[0:5]:
filename = obs[0]
show_image(filename)
from skimage.io import imread
from skimage.color import convert_colorspace, rgb2gray
from helpers import show_image, save_image
# Load image
lena = imread("../lena.jpg")
# Draw and display an image
plt.imshow(lena)
plt.show()
# Matrix indices are [rows, columns]
# Extract the rows 100 to 200 & all columns
sub_lena = lena[100:200, 0:-1]
show_image(sub_lena)
# Accessing color channels: third index -> [rows, cols, channel]
red_lena = lena[:, :, 0]
show_image(red_lena, cmap="gray")
# Convert to grayscale and use different colormap
gray_lena = rgb2gray(lena)
show_image(gray_lena, colormap="inferno")
# Convert to other colorspace
hsv_lena = convert_colorspace(lena, "RGB", "HSV")
# Only show saturation
sat_lena = hsv_lena[:, :, 1]
show_image(sat_lena, "gray")
from skimage.io import imread
from skimage.color import convert_colorspace, rgb2gray
from helpers import show_image, save_image
# Load image
lena = imread("../lena.jpg")
# Draw and display an image
plt.imshow(lena)
plt.show()
# Matrix indices are [rows, columns]
# Extract the rows 100 to 200 & all columns
sub_lena = lena[100:200 , 0:-1]
show_image(sub_lena)
# Accessing color channels: third index -> [rows, cols, channel]
red_lena = lena[:,:,0]
show_image(red_lena, cmap="gray")
# Convert to grayscale and use different colormap
gray_lena = rgb2gray(lena)
show_image(gray_lena, colormap="inferno")
# Convert to other colorspace
hsv_lena = convert_colorspace(lena, "RGB", "HSV")
# Only show saturation
sat_lena = hsv_lena[:,:,1]
show_image(sat_lena, "gray")