[Soc-2016-dev] Weekly Report #11, UV-Tools

[Phil Gosch]

Report #11 for UV-Tools:

As always this can also be found at the wiki: https://wiki.blender.org/index.php/User:SaphireS/gsoc2016/reports#Report_.2311_for_UV-Tools:

What I did this week:

* Discovered a few bugs in packing and margin computation which were introduced by the necessary refactor last week. Took some time until I pinned most of them down, but packing solution computation now works as well as it did before the refactor but now with margin support and cleaner code! A picture showing the margin option can be found here: [1]

* Again reading lots of papers and doing research regarding the best way to add concave hull support. Basically there are 3 methods: Decomposition of the shape into convex shapes, Minkowski sums including cleanup of the resulting shape and a orbital sliding approach which slides one shape around the other by computing sliding vectors and intersections. This paper proved to be very helpful: [2]

After discussion with my mentor we decided the decomposition-method was the best approach in view of the approaching deadline since it makes use of convex hull NFP computation which is already in place and the possibility to skip a recombination step by just checking if points are inside other NFPs. Additionally the decomposition can be prototyped by using triangulation of a shape in combination with a fast "point inside triangle" test.

* Finishing refactor and cleanup of code. Implement smaller tasks marked as "ToDo" in code.

  What I plan on doing next week:

* Continuing implementation of concave hull support.

* Bugfixing and testing.

* Preparing for endterm evaluation.

  Questions:

-

  [1] https://wiki.blender.org/index.php/File:Packing_margin_2.jpg
  [2] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.440.379&rep=rep1&type=pdf


-- 
pixel-pusher at saphirestudio.at