Here, I used morph an image of myself and a friend. I selected corresponding points on each image, then computed the triangulations for each image and their average corresponding points. Using inverse warping and interpolation, I mapped the points from the average image to the original images, obtaining a reshaped image. Then I cross dissolved the two images.
Here, I used the Danish Faces set. I used all males with a neutral expression and a forward profile. I transformed each image's shape to match the average, then cross dissolved all of the images equally.
Here, I create a caricature of one of the Danish men from the previous data set. I decided to use one of the men in the sample, because I do not look similar to them. To find the caricature, I found the difference between the average and the given image, then added that difference to my given image, and transformed to fit that new triangulation shape.
Here, I used the same Danish man from the previous part, and changed his gender. I did this by finding the average female, then transforming the Danish man's face to fit that shape. Then, I cross dissolved.