I am working on scaling a skeleton model to fit the human surface model, like MakeHuman model and SMPL human model.
I attached the experimental markers on the synthetic human model. I also create a static trial trc with a general static pose from those markers.
The names of virtual markers on the skeleton are the same as those on the synthetic human model. I used the scale tool in OpenSim GUI to set up the measurement set. I scaled the skeleton and used Adjust Model Markers for evaluation. After each Preview static pose, I move the virtual markers on the unscaled skeleton, close the scaled model and restart another Preview. I take a long time on the iterative scaling process, but I can't get a well-scaled model.
Did I misunderstand the workflow of scaling? I think I follow the videos and the instruction in https://simtk-confluence.stanford.edu:8 ... th+Scaling... I can't get a correct scale that matches the human surface model....
I also watched the video https://www.youtube.com/watch?v=ZG7wzvQC6eU. I felt a little bit confused, the Adjust Model Markers is in Registration and the Registration is after scaling. The registration is not mentioned in previous videos and instructions. There is a note in the video at 30:33. "Note: realistic time frame for scaling a model from start-to-finish can be 10-12 hours". What is the meaning of the note? What makes the scaling take so long...? Does it include markers placement in the real-world?
I don't have any motion capture machine or any real-world experience in mocap... I am trying to use synthetic data to scale a skeleton to a surface human model...
Please help....