Once we have the coordinates for the independent features, we apply distortions to them (explosions to scale, and implosions to pinch). We then blur the isolated, manipulated features back into the original image in order to create a seamless caricature effect.

Now we can create basic caricatures, but in order to communicate with the machine, we need to take our image and turn it into a set of points that can be sent to our machine. To do this, edge detection is our friend. We first create a blurred, grayscale image of the Caricature, and then apply a Canny edge detection filter to it. From the edge silhouette, we translate the contours into a set of points with x, y coordinates.

This is how we draw Alien Obama!