Wanna know what I'd pay a lot of money for?
WillowRaven
Posts: 3,787
in The Commons
Wanna know what I'd pay a lot of money for? A script or even an associated software that lets me load a photograph that the software can detect the 'bones' and automatically create a pose for v4m4-GMF8 ... that' would be worth gold to me.
Daz 3D is part of 
Connect
DAZ Productions, Inc.
7533 S Center View Ct #4664
West Jordan, UT 84084
Licensing Agreement | Terms of Service | Privacy Policy | EULA
© 2025 Daz Productions Inc. All Rights Reserved.
Comments
Unfortunately, that's beyond what can be achieved at the moment.
Markerless motion tracking does exist, but it relies on a depth camera (or several angles to provide equivalent data), and it also has access to data from surrounding frames.
Possibly someone will create an AI that can do such things in time, but image recognition and analysis is one area where computers still struggle.
The technical term for this is "3D pose estimation".
There are a number of research papers on the subject...
https://en.wikipedia.org/wiki/3D_pose_estimation
https://medium.com/@victoriamazo/3d-human-pose-estimation-ce1259979306
https://ps.is.tuebingen.mpg.de/projects/3d-pose-from-images
https://paperswithcode.com/task/3d-human-pose-estimation
https://nanonets.com/blog/human-pose-estimation-3d-guide/
https://studios.disneyresearch.com/2018/07/18/ar-poser-automatically-augmenting-mobile-pictures-with-digital-avatars-imitating-poses/
...but no Daz Studio specific product that I'm aware of.
Even if I had to 'mark' or target the bones and joints, much like using the pen tool in photoshop, and the script reads my points, that would be a huge timesaver.
Wasn't there a physical, 3D figure that you could pose, and then those poses got translated into a software? I don't remember the name of it, or whether it really was in this lifetime. :D
pretty sure that was what they were doing in like the 90's... so im sure it's a thing.
As a lame alternative, you can use a photo to generate a 3d model, then use that as pose reference.
Working with a 3d pose reference makes the process a lot simpler vs using 2d pose references.
You could try a Kinect and do the pose of the picture yourself for Kinect assuming it's not a somersault in mid flip or that you can't do lovely standing somersaults yourself on land on a dime in a small room.
Doesn't some of the new Houdini 18.5 make this easier? I don't think it is dump photo get motion, but it is pretty interesting. There is a thread here with more info... here is a link I grabbed from it that is pretty interesting if you've go tthe time
https://www.youtube.com/watch?v=zMrvCWy85xM&feature=youtu.be&t=492