Overview

In this assignment we morph my face into someone else's face, compute the mean of a population of faces and extrapolate from a population mean to create a caricature of myself.

Process

This is a rough overview of the process:

1. Align the images
2. Pick correspondence points
3. Average the correspondence points of the different images
4. Calculate the Delaunay triangulation of the average correspondence points
5. Get the transformation to shape each triangle of both images into the triangles from the previous step
6. Cross dissolve the images

Mean Face:

We can use this process to get the mean face of a certain population of people. We took a data set of 40 danish people smiling.

Average the correspondence points of the different images

Calculate the Delaunay triangulation of the average correspondence points

Get the transformation to shape each triangle of ALL images into the triangles from the previous step

Cross dissolve the images

Caricature:

We can use this process and the mean images to get a caricature of my face by exaggerating features of my face while morphing with the mean face. Caricature with increasing values for the features of my face are following:

Bells and Whistles:

Changing Ethnicity:

We can use this process and the mean images to get get a midway image of my face and the average Dane to change my race and make me look more Danish

Bells and Whistles:

Changing Gender:

We can use also use this to morph a person of one gender to have more features of the opposite sex. For this example I take the average female Dane (gotten in the sane way as the average Dane earlier) and morph it with one male Dane at different dissolve and warp factors

As we can see doing a combination of morphing and cross dissolving yields the best results.

Thanks for joining me on this ride. GG