In [ ]:
%matplotlib inline
import numpy as np
import matplotlib.pyplot as plt
In [ ]:
my_matrix = np.array([[1,3,5],[2,5,1],[2,3,8]])
print(my_matrix)
In [ ]:
my_matrix.mean() # mean of the whole matrix
In [ ]:
my_matrix.mean(axis=0) # mean of the columns
In [ ]:
my_matrix.mean(axis=0)[0] # mean of the 0th column
In [ ]:
np.mean(my_matrix, axis=0) # alternative
In [ ]:
my_matrix.mean(axis=1) # mean of the rows
In [ ]:
my_matrix.flatten() # convert to 1D (useful for some plotting)
In [ ]:
plt.style.use('ggplot')
plt.rc('axes', grid=False) # turn off the background grid for images
In [ ]:
plt.imshow(my_matrix, interpolation='nearest', cmap=plt.cm.Blues);
In [ ]:
test_image = np.load("./MyData/test_data.npy") # load in a saved numpy array
In [ ]:
test_image.ndim, test_image.shape, test_image.dtype
In [ ]:
plt.imshow(test_image, cmap=plt.cm.gray);
In [ ]:
print("The minimum value of the image is {0:.2f}".format(test_image.min()))
print("The maximum value of the image is {0:.2f}".format(test_image.max()))
print("The mean value of the image is {0:.2f}".format(test_image.mean()))
print("The standard deviation of the image is {0:.2f}".format(test_image.std()))
In [ ]:
plt.hist(test_image.flatten(),bins=30); #flatten array to get histogram of whole image
In [ ]:
another_test_image = test_image + 8
print("The minimum value of the other image is {0:.2f}".format(another_test_image.min()))
print("The maximum value of the other image is {0:.2f}".format(another_test_image.max()))
print("The mean value of the other image is {0:.2f}".format(another_test_image.mean()))
print("The standard deviation of the other image is {0:.2f}".format(another_test_image.std()))
In [ ]:
plt.imshow(test_image, cmap=plt.cm.gray)
plt.colorbar();
In [ ]:
fig, ax = plt.subplots(1,5,sharey=True)
fig.set_size_inches(12,6)
fig.tight_layout()
ax[0].imshow(test_image, cmap=plt.cm.viridis)
ax[0].set_xlabel('viridis')
ax[1].imshow(test_image, cmap=plt.cm.hot)
ax[1].set_xlabel('hot')
ax[2].imshow(test_image, cmap=plt.cm.magma)
ax[2].set_xlabel('magma')
ax[3].imshow(test_image, cmap=plt.cm.spectral)
ax[3].set_xlabel('spectral')
ax[4].imshow(test_image, cmap=plt.cm.gray)
ax[4].set_xlabel('gray');
In [ ]:
plt.imsave('Splash.png', test_image, cmap=plt.cm.gray) # Write the array I to a PNG file
my_png = plt.imread('Splash.png') # Read in the PNG file
print("The original data has a min = {0:.2f} and a max = {1:.2f}".format(test_image.min(), test_image.max()))
print("The PNG file has a min = {0:.2f} and a max = {1:.2f}".format(my_png.min(), my_png.max()))
In [ ]:
X = np.linspace(-5, 5, 500)
Y = np.linspace(-5, 5, 500)
X, Y = np.meshgrid(X, Y) # turns two 1-d arrays (X, Y) into one 2-d grid
Z = np.sqrt(X**2+Y**2)+np.sin(X**2+Y**2)
Z.min(), Z.max(), Z.mean()
In [ ]:
from matplotlib.colors import LightSource
ls = LightSource(azdeg=0,altdeg=40)
shadedfig = ls.shade(Z,plt.cm.copper)
fig, ax = plt.subplots(1,2)
fig.set_size_inches(8,4)
fig.tight_layout()
ax[0].imshow(shadedfig)
contlevels = [1,2,Z.mean()]
ax[1].imshow(shadedfig)
ax[1].contour(Z,contlevels);
In [ ]:
my_doctor = plt.imread('./MyData/doctor5.png')
print("The image my_doctor has a shape [height,width] of {0}".format(my_doctor.shape))
print("The image my_doctor is made up of data of type {0}".format(my_doctor.dtype))
print("The image my_doctor has a maximum value of {0}".format(my_doctor.max()))
print("The image my_doctor has a minimum value of {0}".format(my_doctor.min()))
In [ ]:
plt.imshow(my_doctor,cmap=plt.cm.gray);
In [ ]:
fig, ax = plt.subplots(1,4)
fig.set_size_inches(12,6)
fig.tight_layout()
# You can show just slices of the image - Rememeber: The origin is the upper left corner
ax[0].imshow(my_doctor, cmap=plt.cm.gray)
ax[0].set_xlabel('Original')
ax[1].imshow(my_doctor[0:300,0:100], cmap=plt.cm.gray)
ax[1].set_xlabel('[0:300,0:100]') # 300 rows, 100 columns
ax[2].imshow(my_doctor[:,0:100], cmap=plt.cm.gray) # ":" = whole range
ax[2].set_xlabel('[:,0:100]') # all rows, 100 columns
ax[3].imshow(my_doctor[:,::-1], cmap=plt.cm.gray);
ax[3].set_xlabel('[:,::-1]') ; # reverse the columns
In [ ]:
fig, ax = plt.subplots(1,2)
fig.set_size_inches(12,6)
fig.tight_layout()
CutLine = 300
ax[0].imshow(my_doctor, cmap=plt.cm.gray)
ax[0].hlines(CutLine, 0, 194, color='b', linewidth=3)
ax[1].plot(my_doctor[CutLine,:], color='b', linewidth=3)
ax[1].set_xlabel("X Value")
ax[1].set_ylabel("Pixel Value")
In [ ]:
from scipy import ndimage
In [ ]:
fig, ax = plt.subplots(1,5)
fig.set_size_inches(14,6)
fig.tight_layout()
ax[0].imshow(my_doctor, cmap=plt.cm.gray)
my_doctor_2 = ndimage.rotate(my_doctor,45,cval=0.75) # cval is the value to set pixels outside of image
ax[1].imshow(my_doctor_2, cmap=plt.cm.gray) # Rotate and reshape
my_doctor_3 = ndimage.rotate(my_doctor,45,reshape=False,cval=0.75) # Rotate and do not reshape
ax[2].imshow(my_doctor_3, cmap=plt.cm.gray)
my_doctor_4 = ndimage.shift(my_doctor,(10,30),cval=0.75) # Shift image
ax[3].imshow(my_doctor_4, cmap=plt.cm.gray)
my_doctor_5 = ndimage.gaussian_filter(my_doctor,5) # Blur image
ax[4].imshow(my_doctor_5, cmap=plt.cm.gray);
ndimage
can do much more: http://scipy-lectures.github.io/advanced/image_processing/
In [ ]:
redfilter = plt.imread("./MyData/sphereR.jpg")
redfilter.shape,redfilter.dtype
In [ ]:
redfilter = plt.imread("./MyData/sphereR.jpg")[:,:,0]
redfilter.shape,redfilter.dtype
In [ ]:
plt.imshow(redfilter,cmap=plt.cm.gray);
In [ ]:
greenfilter = plt.imread("./MyData/sphereG.jpg")[:,:,0]
bluefilter = plt.imread("./MyData/sphereB.jpg")[:,:,0]
In [ ]:
fig, ax = plt.subplots(1,3)
fig.set_size_inches(12,3)
fig.tight_layout()
ax[0].set_title("Red Filter")
ax[1].set_title("Green Filter")
ax[2].set_title("Blue Filter")
ax[0].imshow(redfilter,cmap=plt.cm.gray)
ax[1].imshow(greenfilter,cmap=plt.cm.gray)
ax[2].imshow(bluefilter,cmap=plt.cm.gray);
In [ ]:
rgb = np.zeros((480,640,3),dtype='uint8')
print(rgb.shape, rgb.dtype)
plt.imshow(rgb,cmap=plt.cm.gray);
In [ ]:
rgb[:,:,0] = redfilter
rgb[:,:,1] = greenfilter
rgb[:,:,2] = bluefilter
In [ ]:
fig, ax = plt.subplots(1,4)
fig.set_size_inches(14,3)
fig.tight_layout()
ax[0].set_title("Red Filter")
ax[1].set_title("Green Filter")
ax[2].set_title("Blue Filter")
ax[3].set_title("All Filters Stacked")
ax[0].imshow(redfilter,cmap=plt.cm.gray)
ax[1].imshow(greenfilter,cmap=plt.cm.gray)
ax[2].imshow(bluefilter,cmap=plt.cm.gray)
ax[3].imshow(rgb,cmap=plt.cm.gray);
In [ ]:
print("The image rgb has a shape [height,width] of {0}".format(rgb.shape))
print("The image rgb is made up of data of type {0}".format(rgb.dtype))
print("The image rgb has a maximum value of {0}".format(rgb.max()))
print("The image rgb has a minimum value of {0}".format(rgb.min()))
In [ ]:
rgb[:,:,0] = redfilter * 1.5
plt.imshow(rgb)