OpenSim 3.3 or 3.2 and Scale Tool

[Ana Azevedo]

Hello all,

We are using an OpenSim model to investigate the effects of single-leg landings, and have some questions related to the 1st steps using OpenSim.
1) We start using OpenSim 3.3, but we rolled back to version 3.2, because we were not able to see the virtual and experimental marker sets once we applied the Scale Tool. We cannot identify where the error is shown or where to make any adjustments to the marker set if needed (v3.3). However, we would like to see and identify the errors, and understand how to appropriately adjust them in version 3.3. Could you provide a brief explanation where we can identify the error of the markers and how to adjust them (v3.3), as we did not find any documentation on this matter, please?

2) At which point in the process should we provide a weight to the marker: at scale tool or inverse kinematics tool? We know that we should weight them to give more "relevance" to some markers, but this step is not explicit in the OpenSim documentation.

3) After scaling the model (Scale Tool), what is the error (max and RMS) that we should be looking at, between 2 and 4 cm (max) and 2 cm (RMS)?

4) Where can we find more information concerning the muscles that have more than one part? For example, the gluteus medius (glu-med1, glu-med2, glut-med3). We speculate that it is related with muscles action, but can you guide us to documentation (e.g., OpenSim website, manuscripts, etc.) regarding this topic?

5) While the tutorials are extremely useful, some of them do not entirely match with the current version (v3.3). Do you have an update planned for the tutorials (i.e., the scaling tool and IK)?

Thank you for your time and consideration,
Ana Azevedo

[jimmy d]

1) We start using OpenSim 3.3, but we rolled back to version 3.2, because we were not able to see the virtual and experimental marker sets once we applied the Scale Tool. We cannot identify where the error is shown or where to make any adjustments to the marker set if needed (v3.3). However, we would like to see and identify the errors, and understand how to appropriately adjust them in version 3.3. Could you provide a brief explanation where we can identify the error of the markers and how to adjust them (v3.3), as we did not find any documentation on this matter, please?

You can view the difference between the (scaled and registered) model and the experimental data by previewing the experimental data and removing the display offset of the scaled model. By removing the offset you will overlay the model and the data and be able to visually inspect.

It is not possible to get the errors between experimental and model markers through the GUI. There is a previous post on this; viewtopicPhpbb.php?f=91&t=7965&p=21366&start=0&view=

2) At which point in the process should we provide a weight to the marker: at scale tool or inverse kinematics tool? We know that we should weight them to give more "relevance" to some markers, but this step is not explicit in the OpenSim documentation.

Sorry, but I don't know what you mean by 'this is not explicit in the OpenSim documentation'? There is a whole page on how IK works (https://simtk-confluence.stanford.edu:8 ... tics+Works) and the Scale Tool documentation states that IK is used to pose the model during scaling (https://simtk-confluence.stanford.edu:8 ... rPlacement). There is also a pane in both Tools (Scale and IK) for entering marker weights. Is it not clear to you that you need to specify marker weights in both tools?

To answer your question- you have to provide weights for scale and IK. The wiki documentation talks about weights for both tools (Scale and InverseKinematics). The Scale and IK Tool will need to have different weights since the markers that are important to you will most likely be different between Scale and IK.

3) After scaling the model (Scale Tool), what is the error (max and RMS) that we should be looking at, between 2 and 4 cm (max) and 2 cm (RMS)?

That depends on your model, data and research question. Some markers could be placed on the subject far from the corresponding location on the model, in such a case you would expect large error values. When you assume high accuracy, like a joint center, you would want small differences (<1cm). No same model or experimental trial is the same, and the experimenter can only define the amount of acceptable error for each marker.

4) Where can we find more information concerning the muscles that have more than one part? For example, the gluteus medius (glu-med1, glu-med2, glut-med3). We speculate that it is related with muscles action, but can you guide us to documentation (e.g., OpenSim website, manuscripts, etc.) regarding this topic?

http://nmbl.stanford.edu/wp-content/upl ... al2016.pdf
http://nmbl.stanford.edu/publications/pdf/Hicks2015.pdf

5) While the tutorials are extremely useful, some of them do not entirely match with the current version (v3.3). Do you have an update planned for the tutorials (i.e., the scaling tool and IK)?

What do you mean by 'do not entirely match'? The documentation is being updated for the upcoming 4.0 release. What updates are you hoping to see?

[Wayne Johnson]

James,
Thanks for your response. I do understand that "... you have to provide weights for scale and IK." But is this true for the scale factors as well? This is the 2nd tab in the scale tool. Specifically, do I need to complete scale factors for all bodies prior to running the scaling tool? I would think that the act of running the scale tool would actually generate the scale factors. Yet if I complete the scale factors tab, it does report an applied scale factor even before I actually Run the scale tool.
Best,
Wayne

[jimmy d]

No, you don't have to scale every body in your model. You can scale bodies using either manual factors that you input, or use marker pairs to calculate the scale factor. More detailed information is on the wiki; https://simtk-confluence.stanford.edu:8 ... aleFactors