The requirements for this project can be found on the course website.
For each image, I added 8 set of extra points (the 4 corners and 4 midpoints of the edges).
Then, I calculated the mean of the points of image 1 and image 2 and used this to generate the delaunay triangles (matlab: delaunay).
I used inverse warpping.
For affineCompute, I used the bayercentric coordinate method.
I used 22 images (excluded 20_minjaele [has 44 points] and 05_alex [no points]). For the mean face, I averaged out the points and computed the delaunay triangles as before. For the dissolve, I added all the warpped images and divide by the total number of images. Here are the mean_class_points.txt (added 8 more points as discussed earlier).