Latest Page Update: 23-09-2026

Exercise 7 - Geometric transformations and landmark based registration¶

In this exercise, we will explore geometric transformations of images and landmark based registration.

Learning Objectives¶

After completing this exercise, the student should be able to do the following:

Use skimage.transform.rotate to rotate an image using different rotation centers, different background filling strategies (constant, reflection, warping) and automatic scaling of the output image.
Construct an Euclidean (translation plus rotation) transform using skimage.transform.EuclideanTransform.
Apply a given transform to an image using skimage.transform.warp.
Compute and apply the inverse of a transform.
Construct a similarity (translation, rotation plus scale) transform using skimage.transform.SimilarityTransform.
Use the skimage.transform.swirlto transform images.
Compute and visualize the blend of two images.
Manually place landmarks on an image.
Visualize sets of landmarks on images.
Compute the objective function \(F\) between two sets of landmarks.
Use the estimate function to estimate the optimal transformation between two sets of landmarks.
Use the skimage.transform.matrix_transform to transform a set of landmarks.
Implement and test a program that can transform and visualize images from a video stream.

Installing Python packages¶

In this exercise, we will be using both scikit-image and OpenCV. You should have these libraries installed, else instructions can be found in the previous exercises.

We will use the virtual environment from the previous exercise (course02503).

Exercise data and material¶

Download the data you'll need for this exercise by clicking here: Data

Alternatively, you can also fetch the data for the whole course through the Image Analysis GitHub repository. See the Data and GitHub section for more information. If you're using Git, it may be wise run:

git pull

As this will fetch updates to the material, which may happen throughout the course.