The below is the correspondence points that I chose to be significant points of translation. By choosing the points in the same order, I can guarantee that the Delaunay transformation calculated on the mid-way shape should give me the same result for the points of both images.
I computed the Mid-Way face by first morphing both images into the mean shape, and then cross dissolving the pixel values from each image to get the final result. Computations for this section included setting correspondance points for each image and creating a Delaunay triangulation, calculating the Affine transformation matrix from one triangle to the other, and then morphing each triangle. The result is the exact middle image, both in appearance and in shape, between the two original images.
The following is a gif of ~45 frames of weighted morphing that favors one image's appearance and shape and over time ends up favoring the other image, creating a nice smooth morphing effect.
This portion of the project consisted of finding the average face over a subset of the danish population, thus finding the "mean face" of a danish person. I just found the average of all the images' shapes and appearances, first morphing each image to the average image's shape and then adding a weighted version of each pixel to the returned image. Below are some of the images morphed to the average image's shape and the mean face of the danish population. Additionally, the next two images are my face morphed to the shape of the average danish person, and then the average danish person's face morphed to my face's shape.
The image below is a caricature of my face I acheived by extrapolating from the danish population mean I calculated in the previous portion of the project. Instead of seeing myself as more danish, the image below exagerrates the portion of my face that are not similar to the Danish population mean, which in turn makes me "less Danish".
This part of the project consisted of taking a picture of your friend and, in my case, changing their ethnicity. The two first pictures are a picture of my friend, and the face of the average black man. The first image is the result of only blending the appearances, which you receive from a naive cross-dissolve, and the second image is the result of only blending the shapes of the two images. Finally, the last image was found by using the same techniques in previous portions of my project, creating an image of how my friend would look if he were black.