search

Tomoyon / makehuman

Automatically exported from code.google.com/p/makehuman
0 stars 0 forks source link

idea for the character poses #113

Closed GoogleCodeExporter closed 9 years ago

GoogleCodeExporter commented 9 years ago 
Hello,
My name is Jeremy Barnabe, Belgian graphic designer.

I have an idea about the poses of characters. My English is not very good, I'll 
explain it by 3 sketch:
http://data.imagup.com/member/1121429380_V1.jpg
http://data.imagup.com/member/1121429391_V2.jpg
http://data.imagup.com/member/1121429400_V3.jpg

the idea is that there is only one poses a configuration of red dot.
In the same way with your mobile phone, there is only one word if your T9 
dictionary is enabled.

The algorithm should infer the pose according to:
Red dots placed
Of movement impossible (wrist below the elbow,...)

What do you think?

Jeremy Barnaby
Delphi_jb@yahoo.fr

Original issue reported on code.google.com by camera.d...@gmail.com on 30 May 2011 at 2:08

Attachments:

GoogleCodeExporter commented 9 years ago 
Not bad. A very interesting idea.
It need a large library of poses, but I like it.
Thanks!

Original comment by mbasti...@gmail.com on 30 May 2011 at 9:53

GoogleCodeExporter commented 9 years ago 
You have 2 options for this feature:
* a large library of installation (a very large then!)
* reasoning algorithm that I call "inference by the retrospective space"

What is it?

you start by deducing the position of the hips. IF your red dots allow two 
spatial possibilities, then the two possibilities are registered (the software 
will choose one at random). Then you place the red dot below (eg, knees)
the two possibilities may occur:

1 - from the position of the hip, knee position is possible. is validated
2 - from the position of the hip, knee position is impossible.

If you are in the second case, then the software will "backfire" on the hip, 
remember the alternative space and use it. Then, the knee will be good from the 
hip.

If, starting from the hip to the foot, several options are possible, then the 
software will submit them to the user to choose.

Sincerely,

Delphi_JB

Original comment by camera.d...@gmail.com on 31 May 2011 at 8:50

GoogleCodeExporter commented 9 years ago 
Nice idea, but position possibility determination algorithm should handle 
situations like this (watch attachment)

Original comment by root.ta...@gmail.com on 29 Aug 2011 at 4:59

Attachments:

GoogleCodeExporter commented 9 years ago 
Yes indeed, perhaps more should be a check options such as:

contortionist:
* Yes * No

The algorithm would have a lower tolerance of some movement, such as the 
curvature of the spine.

I discussed this idea with other colleagues in computer graphics, specializing, 
as I am, in 3D, and all very promptly said it was a good idea, we should use it 
... ^ ^

Original comment by camera.d...@gmail.com on 29 Aug 2011 at 5:29

GoogleCodeExporter commented 9 years ago 
I think it's a good and innovative idea, also.
It's frustrating for me to see it and haven't enough resources to work on it. A 
pose system like this require at least a coder dedicated to it only. If anyone 
know someone that can help, please ask him. I'm available to discuss the 
implementation.

Best,

Manuel

Original comment by mbasti...@gmail.com on 29 Aug 2011 at 5:48

GoogleCodeExporter commented 9 years ago 
I will look to find someone but that will not be easy. I already asked the 
person most qualified for this type of coding but it has its own projects...

Original comment by camera.d...@gmail.com on 29 Aug 2011 at 7:50

GoogleCodeExporter commented 9 years ago

Original comment by mflerack...@gmail.com on 29 Dec 2011 at 1:08

GoogleCodeExporter commented 9 years ago 
Manuel, Is this still the sort of direction  you'd like to head with the poses? 
Is this a desired work flow: Image of human model(real or otherwise)used to 
"set" the pose of the MH model?

Original comment by supu...@gmail.com on 1 Jun 2012 at 6:16

GoogleCodeExporter commented 9 years ago 
Excuse me, I do not quite understand your question...
if it's for the head, I think the direction of it will be in a special feature, 
may be in a second window with a zoom on It.

Original comment by camera.d...@gmail.com on 1 Jun 2012 at 9:10

GoogleCodeExporter commented 9 years ago 
While we are not immediately following this approach, I can announce that we 
are working on a static pose system.
Dynamic posing is also something you can expect in the future.

Original comment by jonas.ha...@gmail.com on 14 Nov 2012 at 1:25

GoogleCodeExporter commented 9 years ago 
Moved: http://www.makehuman.org/bugtracker/issues/86

Original comment by mbasti...@gmail.com on 11 Feb 2014 at 8:34

GoogleCodeExporter commented 9 years ago

Original comment by mbasti...@gmail.com on 11 Feb 2014 at 8:34