Open venation pattern

Comment

Comment by martyn hogg on September 4, 2016 at 9:32am

I posted a question about it in the Anenome group forum but no one replied :(

Anenome Leaf venation attempt

I gave up on working out the direction vectors based on food points and just randomised the directions of growth to keep the definition simpler.

Have a look if you have time, but don't worry too much... I need to read that paper and start again really!

Comment by Daniel González Abalde on September 4, 2016 at 6:38am

I understand, is convenient to do it from code. If you post the specific problem in the forum can be easy to resolve from code. But why you can not add a branch?

Here I use Hoopsnake because it is uncomfortable for me to use Anemone in experimental definitions, if something goes wrong (really wrong) may take too long time because it automatically runs the whole number of cycles. Hoopsnake instead allows me to understand what happens in each iteration, making it much easier to find and correct errors. However, none is perfect, I have to use an anonymous type of VB (with very few lines of code) to move -as a trailer- all the data from end to beginning of the loop.

Comment by martyn hogg on September 4, 2016 at 6:17am

is there a reason why you use Hoopsnake not Anenome?

I got stuck with Anenome because I couldn't get it to add a new data tree branch with each iteration. My data just became a mess.

Comment by Daniel González Abalde on September 3, 2016 at 1:04pm

Is part of the process. When a new branch arises, is set the half of radius of the predecessor branch at that point, and when a branch extends, a proportion of the last radius is added at the end of the branch.
I have tried various methods, including transportation through every vein, but for now this is the best looking.
I have not explored too much, but my idea is to establish an equation as a parameter to set/modify any value (radio, amount of water, light, food from the source points...) with a transport method, which is just a road of indices structured in datatree, that reference from the start node to all ends of the branches, or opposite direction.

Comment by martyn hogg on September 3, 2016 at 12:44pm

Looks great! How do you deal with thickening the branches? Is this just done as a function of the distance from the start point?

You do it on the lines in the original image too... is that a post processing thing or is it part of the generation process?

Comment by Daniel González Abalde on September 3, 2016 at 11:39am

Custom wrap + wbLoop (L=1)

Comment by David Stasiuk on September 3, 2016 at 10:50am

the closed is tricky

Comment by martyn hogg on September 3, 2016 at 4:20am

I was trying to apply the pattern to a mesh surface so I could CNC machine the pattern into wood. That way i wouldn't need the structural strength.

When I get time I'll read that paper properly and try again!

Comment by Daniel González Abalde on September 3, 2016 at 4:08am

Thanks Martyn.

Yeap, it is necessary to give some topology. 
A few days ago I did this only with VB, and by failing to make the closed pattern, I am now doing with Hoopsnake+VB to better understand what goes wrong.

I hope to achieve close it because the closed pattern is needed for structural resistance (if you want to print).

Comment by martyn hogg on September 3, 2016 at 3:59am

Really cool!

I tried this with anenome but couldn't really control it properly.

Are you going to try the closed venetion pattern too?