eth-igl / gp2024-Assignments

0 stars 0 forks source link

Geometric Handle Selection (Task 4) #48

Closed 55mmm closed 3 weeks ago

55mmm commented 3 weeks ago

Hello,

I have two questions about geometric handle selection in task 4.

  1. I believe it is not guaranteed that geometric handle selection results in strictly disjoint sets in the sense of the third bullet point in Task 1.1, as opposed to the provided reference handles. Can I keep vertices in intersecting handle sets? Or can I assume that vertices which would be bound to more than one handle should be removed from all handles?
  2. The handout indicates a factor of 1.5. In my personal experience - at least for the hand model - other, slightly larger factors provide much better-looking results. Is there some room for negotiation here, or is that a strict 1.5?

Best regards,

Michael

segaviv commented 3 weeks ago

Hi,

  1. The handles should indeed be strictly disjoint. You can experiment and decide which way is the best to handle the situation you mentioned (e.g. dropping the overlapped handles from one of the bones/ splitting them between the bones/ removing them from the handles).
  2. Yes, the suggestion is just an example. You should experiment with different parameters/strategies and find the one that works best.