Previous | Next --- Slide 36 of 47
Back to Lecture Thumbnails
hubbahubba

This seems like a "greedy" algorithm. Are there certain scenarios that it fails to give a good result?

Azure

I don't see why this should work. If you collapse an edge, you change the planes of all the triangles incident on the original two vertices, so shouldn't you have to recompute all those K, and reupdate all of the K_i of the adjacent vertices, etc.? I feel like you wouldn't be able to just reassign K_ij to the new vertex.