General clarifications, simplifications and figure corrections. Using imshow from matplotlib instead of skimage (deprecated). Setting default colormap.

Author: mwhamgenomics

episodes/01-opening-and-checking-an-image.md

@@ -51,15 +51,18 @@ Python has loaded and stored the image as a Numpy `array` object of numbers, and
 to it, which is why we get a matrix of numbers printed to the screen.
 
 If we want to see what the image looks like, we need tell Python to display it
-as an image. We can do this with the `imshow` function from skimage, which can
-be imported and called in a similar way to `imread`:
+as an image. We can do this with the `imshow` function from matplotlib, which you
+may already be familiar with as a library for drawing plots and graphs, but it can
+also display images.
 
 ```python
-from skimage.io import imshow
-imshow(image)
+import matplotlib.pyplot as plt
+
+plt.set_cmap('gray')  # by default, single-channel images will now be displayed in greyscale
+plt.imshow(image)
 ```
 
-You should see the image displayed below the current cell:
+You should now see the image displayed below the current cell:
 
 ![](fig/1_1_imshow.jpg){alt='Displaying an image with imshow'}
 
@@ -79,7 +82,7 @@ are the X and Y axes. The third axis in most cases will represent a number of ch
 We can select a single channel by **indexing** the array:
 
 ```python
-imshow(image[:, :, 2])
+plt.imshow(image[:, :, 2])
 ```
 
 Here, we select the entire X and Y axes using `:` with no numbers around them, and
@@ -121,7 +124,7 @@ Next we can use `.T` to return a **transposed** version of the image. Running
 
 ```python
 channel = image[:, :, 1]
-imshow(channel.T)
+plt.imshow(channel.T)
 ```
 
 ![](fig/1_2_transpose.jpg){alt='Transposed image'}
@@ -174,25 +177,22 @@ Running `image.nbytes` shows that it takes up 786432 bytes, or ~786 kilobytes, o
 ## Displaying one channel at a time
 
 We've seen from exercise 1 that we can view single channels by indexing the array. We can also show all
-channels together using matplotlib. You may already be familiar with matplotlib as a library for drawing
-plots and graphs, but it can also display images.
-
-We can diaply a group of images together using a matplotlib figure:
+channels together using a matplotlib figure:
 
 ```python
-import matplotlib
-matplotlib.pyplot.figure(figsize=(12, 6))  # figure size, in inches
+import matplotlib.pyplot as plt
+plt.figure(figsize=(12, 6))  # figure size, in inches
 nchannels = 3
 
 for i in range(nchannels):
     # Use subplot() to create a multi-image figure with 1 row and 3 columns. We need to increment i by 1
     # because range() counts from 0 but subplot() assumes you're counting from 1.
-    matplotlib.pyplot.subplot(1, nchannels, i+1)
-    matplotlib.pyplot.imshow(image[:, :, i])
-    matplotlib.pyplot.title('Channel %s' % i)  # add a plot title
-    matplotlib.pyplot.axis(False)              # we just want to show the image, so turn off the axis labels
+    plt.subplot(1, nchannels, i+1)
+    plt.imshow(image[:, :, i])
+    plt.title('Channel %s' % i)  # add a plot title
+    plt.axis(False)              # we just want to show the image, so turn off the axis labels
 
-matplotlib.pyplot.show()
+plt.show()
 ```
 
 ![](fig/1_4_all_channels.jpg){alt='All image channels'}
@@ -204,7 +204,7 @@ Another useful metric in image analysis is an image's histogram. This can be plo
 the image and passing it to matplotlib:
 
 ```python
-matplotlib.pyplot.hist(image[:, :, 0].flatten(), bins=256)
+plt.hist(image[:, :, 0].flatten(), bins=256)
 ```
 
 First, we need to select a single channel - since different channels may represent different cell organelles
@@ -226,26 +226,26 @@ Starting with displaying a single histogram for one channel:
 ```python
 channel_idx = 0
 channel = image[:, :, channel_idx]
-matplotlib.pyplot.hist(channel.flatten(), bins=255)
-matplotlib.pyplot.title('Channel %s' % channel_idx)
-matplotlib.pyplot.show()
+plt.hist(channel.flatten(), bins=255)
+plt.title('Channel %s' % channel_idx)
+plt.show()
 ```
 
 We could call this three times, each with a different value for `channel_idx`, or we can use
 a **for loop**:
 
 ```python
-matplotlib.pyplot.figure(figsize=(12, 6))  # figure size, in inches
+plt.figure(figsize=(12, 6))  # figure size, in inches
 nchannels = image.shape[-1]
 
 for i in range(nchannels):
     channel = image[:, :, i]
 
-    matplotlib.pyplot.subplot(1, nchannels, i+1)
-    matplotlib.pyplot.hist(channel.flatten(), bins=255)
-    matplotlib.pyplot.title('Channel %s' % i)  # add a plot title
+    plt.subplot(1, nchannels, i+1)
+    plt.hist(channel.flatten(), bins=255)
+    plt.title('Channel %s' % i)  # add a plot title
 
-matplotlib.pyplot.show()
+plt.show()
 ```
 ![](fig/1_5_histograms.jpg){alt='All image channels'}
 
@@ -314,7 +314,7 @@ the X and Y axes, while the second number looks like a number of colour channels
 looks like either a time series or a Z axis. We can show a single frame with:
 
 ```python
-imshow(image[0, 0, :, :])
+plt.imshow(image[0, 0, :, :])
 ```
 
 ![](fig/1_6_nd2_image_frame.jpg){alt='ND2 image'}
@@ -337,14 +337,15 @@ The `imshow()` function can take extra arguments in addition to the image to dis
 of these is called `cmap`, which can apply alternate lookup tables (a.k.a. colour maps):
 
 ```python
-imshow(image[0, 0, :, :], cmap='gray')
+plt.imshow(image[0, 0, :, :], cmap='viridis')
 ```
 
 Skimage uses lookup tables from the plotting library matplotlib. A list of available tables
 can be obtained with:
 
 ```python
-print(sorted(matplotlib.colormaps))
+from matplotlib import colormaps
+print(sorted(colormaps))
 ```
 
 ::::::::::::::::::::::::::::::::::::: challenge
@@ -361,16 +362,16 @@ is one possible solution:
 ```python
 # you'll need to run this again if you overwrote your `image` variable
 image = imread('data/FluorescentCells_3channel.tif')
-matplotlib.pyplot.figure(figsize=(12, 6))
+plt.figure(figsize=(12, 6))
 
-matplotlib.pyplot.subplot(1, 3, 1)
-imshow(image[:, :, 0], cmap='Blues')
+plt.subplot(1, 3, 1)
+plt.imshow(image[:, :, 0], cmap='Blues')
 
-matplotlib.pyplot.subplot(1, 3, 2)
-imshow(image[:, :, 1], cmap='Oranges')
+plt.subplot(1, 3, 2)
+plt.imshow(image[:, :, 1], cmap='Oranges')
 
-matplotlib.pyplot.subplot(1, 3, 3)
-imshow(image[:, :, 2], cmap='YlOrBr')
+plt.subplot(1, 3, 3)
+plt.imshow(image[:, :, 2], cmap='YlOrBr')
 ```
 
 ![](fig/1_7_colormaps.jpg){alt='ND2 image'}

episodes/02-applying-filters.md

@@ -45,14 +45,19 @@ kernel = square(3)  # create a 3x3 square kernel
 There are other shapes of kernel that can be used, and are documented
 [here](https://scikit-image.org/docs/stable/api/skimage.morphology.html).
 
-(Note that as of skimage 0.25.0, the `square` function has been deprecated in
-favour of a new function, [footprint_rectangle](https://scikit-image.org/docs/stable/api/skimage.morphology.html#skimage.morphology.footprint_rectangle).)
+Note that as of skimage 0.25.0, the `square` function has been deprecated in
+favour of a new function, [footprint_rectangle](https://scikit-image.org/docs/stable/api/skimage.morphology.html#skimage.morphology.footprint_rectangle):
+
+```python
+from skimage.morphology import footprint_rectangle
+kernel = square((3, 3))  # also a 3x3 square kernel
+```
 
 ::::::::::::::::::::::::::::::::::::: challenge 
 ## Exercise 6: Applying filters
 
-Look at the documentation pages for the mean and Gaussian filters above. Load the second channel
-of the 30th frame of the test image [skimage.data.cells3d()](https://scikit-image.org/docs/stable/api/skimage.data.html#skimage.data.cells3d):
+Look at the documentation pages for the mean and Gaussian filters above. Load a frame from the second channel
+of the test image [skimage.data.cells3d](https://scikit-image.org/docs/stable/api/skimage.data.html#skimage.data.cells3d):
 
 ```python
 from skimage.data import cells3d
@@ -75,30 +80,30 @@ How do the different methods compare?
 from skimage.data import cells3d
 from skimage.filters import gaussian
 from skimage.filters.rank import mean
-from skimage.morphology import square
+from skimage.morphology.footprints import square
 
 image = cells3d()[30, 1, :, :]
-matplotlib.pyplot.figure(figsize=(10, 12))
+plt.figure(figsize=(10, 12))
 
-matplotlib.pyplot.subplot(3, 2, 1)
-imshow(image)
-matplotlib.pyplot.title('Original')
+plt.subplot(3, 2, 1)
+plt.imshow(image)
+plt.title('Original')
 
-matplotlib.pyplot.subplot(3, 2, 2)
-imshow(mean(image, footprint=square(3)))
-matplotlib.pyplot.title('3x3 mean filter')
+plt.subplot(3, 2, 2)
+plt.imshow(mean(image, footprint=square(3)))
+plt.title('3x3 mean filter')
 
-matplotlib.pyplot.subplot(3, 2, 3)
-imshow(mean(image, footprint=square(9)))
-matplotlib.pyplot.title('9x9 mean filter')
+plt.subplot(3, 2, 3)
+plt.imshow(mean(image, footprint=square(9)))
+plt.title('9x9 mean filter')
 
-matplotlib.pyplot.subplot(3, 2, 4)
-imshow(gaussian(image, sigma=1))
-matplotlib.pyplot.title('Gaussian blur, σ=1')
+plt.subplot(3, 2, 4)
+plt.imshow(gaussian(image, sigma=1))
+plt.title('Gaussian blur, σ=1')
 
-matplotlib.pyplot.subplot(3, 2, 5)
-imshow(gaussian(image, sigma=5))
-matplotlib.pyplot.title('Gaussian blur, σ=5')
+plt.subplot(3, 2, 5)
+plt.imshow(gaussian(image, sigma=5))
+plt.title('Gaussian blur, σ=5')
 ```
 
 ![](fig/2_1_filters.jpg){alt='Filters'}
@@ -131,46 +136,34 @@ and display:
 - image with a rolling ball of radius 50 applied
 - the image with the radius=50 rolling ball subtracted from it
 
-How long does it take to compute?
-
 :::::::::::::::::::::::: solution 
 
 Compute time can be found using Python's datetime library:
 
 ```python
-import datetime
 from skimage.data import coins
 from skimage.restoration import rolling_ball
 
 image = coins()
-matplotlib.pyplot.figure(figsize=(10, 12))
+plt.figure(figsize=(10, 8))
 
-t0 = datetime.datetime.now()
-matplotlib.pyplot.subplot(2, 2, 1)
-imshow(image)
-matplotlib.pyplot.title('Original')
+plt.subplot(2, 2, 1)
+plt.imshow(image, cmap='gray')
+plt.title('Original')
 
-t1 = datetime.datetime.now()
-matplotlib.pyplot.subplot(2, 2, 2)
+plt.subplot(2, 2, 2)
 rolling_ball_100 = rolling_ball(image, radius=100)
-imshow(rolling_ball_100)
-matplotlib.pyplot.title('Rolling ball, r=100')
+plt.imshow(rolling_ball_100, cmap='gray')
+plt.title('Rolling ball, r=100')
 
-t2 = datetime.datetime.now()
-matplotlib.pyplot.subplot(2, 2, 3)
+plt.subplot(2, 2, 3)
 rolling_ball_50 = rolling_ball(image, radius=50)
-imshow(rolling_ball_50)
-matplotlib.pyplot.title('Rolling ball, r=50')
-
-t3 = datetime.datetime.now()
-matplotlib.pyplot.subplot(2, 2, 4)
-imshow(image - rolling_ball_50)
-matplotlib.pyplot.title('Original - rolling ball r50')
-
-print('Original image:', t1 - t0)
-print('Rolling ball 100:', t2 - t1)
-print('Rolling ball 50:', t3 - t2)
-print('Total:', t3 - t0)
+plt.imshow(rolling_ball_50, cmap='gray')
+plt.title('Rolling ball, r=50')
+
+plt.subplot(2, 2, 4)
+plt.imshow(image - rolling_ball_50)
+plt.title('Original - rolling ball r50')
 ```
 
 ![](fig/2_2_rolling_ball.jpg){alt='Rolling ball'}
@@ -195,7 +188,7 @@ filter. This eliminates the need to manually select two sigma values as with Dif
 ::::::::::::::::::::::::::::::::::::: challenge
 ## Exercise 8: Dogs and logs
 
-Load the second channel of the 30th frame of [skimage.data.cells3d()](https://scikit-image.org/docs/stable/api/skimage.data.html#skimage.data.cells3d)
+Load a frame from the second channel of [skimage.data.cells3d()](https://scikit-image.org/docs/stable/api/skimage.data.html#skimage.data.cells3d)
 again:
 
 ```python
@@ -216,22 +209,22 @@ Display a figure of:
 
 ```python
 from skimage.data import cells3d
-from skimage.filters import difference_of_gaussians, laplace
+from skimage.filters import gaussian, difference_of_gaussians, laplace
 
 image = cells3d()[30, 1, :, :]
-matplotlib.pyplot.figure(figsize=(10, 12))
+plt.figure(figsize=(10, 12))
 
-matplotlib.pyplot.subplot(2, 2, 1)
+plt.subplot(2, 2, 1)
 imshow(image)
-matplotlib.pyplot.title('Original')
+plt.title('Original')
 
-matplotlib.pyplot.subplot(2, 2, 2)
+plt.subplot(2, 2, 2)
 imshow(difference_of_gaussians(image, 1, 2), cmap='gray')
-matplotlib.pyplot.title('Difference of Gaussians')
+plt.title('Difference of Gaussians')
 
-matplotlib.pyplot.subplot(2, 2, 3)
+plt.subplot(2, 2, 3)
 imshow(laplace(gaussian(image, sigma=2)), cmap='gray')
-matplotlib.pyplot.title('Laplacian of Gaussian')
+plt.title('Laplacian of Gaussian')
 ```
 
 ![](fig/2_3_dogs_and_logs.jpg){alt='Dogs and logs'}