To create the mid way face, we first have to plot a set of correspondance points on our two images. We do this manually and make sure that we select the points in the same order. Next we compute the average each pair of points and then get a Delaunay triangulation from the averaged points. We then apply the triangulation to the points of each image and the average points. Next, we find an affine tranformation matrix that morphs the triangles of each image to the average triangulation.
The morph sequence was created from a series of mid-way faces with different weights from 0-1 in 1/60 intervals on the facial shape morph and the image color.
I used the Danes facial dataset and used front facing male images. To create the mean fae of a population we take a series of images and their correspondance points and create and average of all correspondance points to calculate the average face shape. Then we morph each face to the average shape and also take and average of color intensity at each pixel to create the average face of a dataset
The Danes dataset only contains annotations along the jaw, the nose, the eyes, and the eyebrows. Because some areas are not annotated, namely the upper head and the neck, morphing a face to the average shape often causes distorted morphings in those areas.
I created caricatures of myself by comparing my face to the average Danish face. We find the difference between my face and the average Danish face and instead of averaging the faces, we extropolate the differences instead to exaggerate the features that make me, me.
According to the caricatures, I have thick and long eyebrows and also a wide jaw.
The I chose to change my ethnicity and gender. I took my photo and morphed the shape and color to the average face of Vogue cover faces. I produced the shape morph by morphing my face to the average Vogue face shape. I made the color morph by morphing the Vogue face to my face shape and then averaging the colors. I also produced the gif that shows the morphing and can get the mid-way face as the middle intermediary face.