Instructor Notes

Estimated Timings

This is a relatively new curriculum. The estimated timings for each episode are based on limited experience and should be taken as a rough guide only. If you teach the curriculum, the Maintainers would be delighted to receive feedback with information about the time that was required for teaching and exercises in each episode of your workshop.

Please open an issue on the repository to share your experience with the lesson Maintainers.

Working with Jupyter notebooks

  • This lesson is designed to be taught using Jupyter notebooks. We recommend that instructors guide learners to create a new Jupyter notebook for each episode.

  • Python import statements typically appear in the first code block near the top of each episode. In some cases, the purpose of specific libraries is briefly explained as part of the exercises.

  • The possibility of executing the code cells in a notebook in arbitrary order can cause confusion. Using the “restart kernel and run all cells” feature is one way to accomplish linear execution of the notebook and may help locate and identify coding issues.

  • Many episodes in this lesson load image files from disk. To avoid name clashes in episodes that load multiple image files, we have used unique variable names (instead of generic names such as image or img). When copying code snippets between exercises, the variable names may have to be changed. The maintainers are keen to receive feedback on whether this convention proves practical in workshops.

Working with imageio and skimage

  • imageio.v3 allows to load images in different modes by passing the mode= argument to imread(). Depending on the image file and mode, the dtype of the resulting Numpy array can be different (e.g., dtype('uint8') or dtype('float64'). In the lesson, skimage.util.img_as_ubyte() and skimage.util.img_as_float() are used to convert the data type when necessary.

  • Some skimage functions implicitly convert the pixel values to floating-point numbers. Several callout boxes have been added throughout the lesson to raise awareness, but this may still prompt questions from learners.

  • In certain situations, imread() returns a read-only array. This depends on the image file type and on the backend (e.g., Pillow). If a read-only error is encountered, image = np.array(image) can be used to create a writable copy of the array before manipulating its pixel values.

  • Be aware that learners might get surprising results in the Keeping only low intensity pixels exercise, if plt.imshow is called without the vmax parameter. A detailed explanation is given in the Plotting single channel images (cmap, vmin, vmax) callout box.

Questions from Learners

Q: Where would I find out that coordinates are x,y not r,c?

A: In an image viewer, hover your cursor over top-left (origin) the move down and see which number increases.

Q: Why does saving the image take such a long time? (skimage-images/saving images PNG example)

A: It is a large image.

Q: Are the coordinates represented x,y or r,c in the code (e.g. in array.shape)?

A: Always r,c with numpy arrays, unless clearly specified otherwise - only represented x,y when image is displayed by a viewer. Take home is don’t rely on it - always check!

Q: What if I want to increase size? How does skimage upsample? (image resizing)

A: When resizing or rescaling an image, skimage performs interpolation to up-size or down-size the image. Technically, this is done by fitting a spline function to the image data. The spline function is based on the intensity values in the original image and can be used to approximate the intensity at any given coordinate in the resized/rescaled image. Note that the intensity values in the new image are an approximation of the original values but should not be treated as the actual, observed data. skimage.transform.resize has a number of optional parameters that allow the user to control, e.g., the order of the spline interpolation. The scikit-image documentation provides additional information on other parameters.

Q: Why are some lines missing from the sudoku image when it is displayed inline in a Jupyter Notebook? (skimage-images/low intensity pixels exercise)

A: They are actually present in image but not shown due to interpolation.

Q: Does blurring take values of pixels already blurred, or is blurring done on original pixel values only?

A: Blurring is done on original pixel values only.

Q: Can you blur while retaining edges?

A: Yes, many different filters/kernels exist, some of which are designed to be edge-preserving.

Troubleshooting

Learners reported a problem on some operating systems, that Shift+Enter is prevented from running a cell in Jupyter when the caps lock key is active.

Introduction

Image Basics

Working with scikit-image

Drawing and Bitwise Operations

Creating Histograms

Blurring Images

Terminology about image boundaries

Take care to avoid mixing up the term “edge” to describe the edges of objects within an image and the outer boundaries of the images themselves. Lack of a clear distinction here may be confusing for learners.


Thresholding

Instructor Note

The histogram of the maize root image may prompt questions from learners about the interpretation of the peaks and the broader region around 0.6. The focus here is on the separation of background and foreground pixel values. We note that Otsu’s method does not work well for the image with the shapes used earlier in this episode, as the foreground pixel values are more distributed. These examples could be augmented with a discussion of unimodal, bimodal, and multimodal histograms. While these points can lead to fruitful considerations, the text in this episode attempts to reduce cognitive load and deliberately simplifies the discussion.


Connected Component Analysis

Capstone Challenge