Using the labeling tool provided in the project spec (https://cal-cs180.github.io/fa23/hw/proj3/tool.html), I created corresponding paris of points on the two images I am trying to merge.
1. Find the Delaunay on the average points.
2. Warp face 1 to the average points.
3. Warp face 2 to the average points.
4. Take the average of the two results.
For those who are curious, these are the intermediate steps.
1. Find the Delaunay on the average points.
2. Define interpolated points given warp_frac.
2. Warp face 1 to the interpolated points.
3. Warp face 2 to the interpolated points.
4. Take the average of the two results according to dissolve_frac.
I am taking the mean face of many man from the FEI Face Database (https://fei.edu.br/~cet/facedatabase.html).
1. Find the Delaunay on the average of all points.
2. Warp all the faces to these average points.
3. Take the average of all results.
Do note that each of the man images I got corresponding points provided in the dataset.
Instead on interpolating, now we are exterpolating, creating geometry that isn't an average of the points. I set the alpha value to be -0.5 and 1.5 to see the effects on exterpolation.
Morphing myself with average of another gender and race (In this case Caucasian Women).