Opensim 4.0 does not complete scaling process.

[Laurent Hendrickx]

Hi all,

I was wondering whether anybody else is experiencing problems with scaling in opensim 4.0

The scaling process seems to get stuck on step 3 "placing markers on mode", it says "executing scaling" but has never completed the process.

I use the 3Dgaitmodel2392 and my own designed markerset, and a static trial that was captured in vicon but converted with MotoNMS to .trc

In OpenSim 3.3 scaling with these files works just fine.
However, since I find OpenSim 4.0 way more userfriendly, I definitely prefer to work in 4.0 if possible.

Is there anybody that might have a solution?

Thanks!

Laurent

[Ayman Habib]

Dear Laurent,

The scaling process had fundamentally stayed the same between 3.3 and 4.0, the only thing that may have changed is the naming convention/uniqueness requirements. If you load the setup files in the GUI and you don't see errors or undefined scale factors/red text-boxes you should be good to go.
It's also possible that there're permission issues. The distributed Models/Tutorials already exercise scaling so the issue is likely model/data-files specific.

Please let us know what you find out,
-Ayman

[Maria Fox]

Hello,

I seem to be having the same issue. I've tried scaling with both the Rajagopal 2016 and Lai 2017 full body models in OpenSim 3.3 and 4.0. The exact same scaling setup works in 3.3 but gets stuck placing markers on model in "executing scaling..." with 4.0. There are no obvious errors during steps 1 and 2 of scaling in 4.0.

I wonder if it could be related to an odd issue I'm having where the hat_ribs_scap.vtp for the two models will not load in 4.0? It originally loaded in one model, but it no longer loads and I get an error message stating:

Code: Select all

Visualizer couldn't open hat_ribs_scap.vtp because:
SimTK Exception thrown at polygonalmesh.cpp:411:
  Error detected by Simbody method PolygonalMesh::loadVtpFile(): Attempt to load a VTK PolyData (.vtp) file from file name 'c:\Users\Maria\Box Sync\Fox Dissertation\OpenSim\Model_Rajagopal2016\Geometry\hat_ribs_scap.vtp' failed with message:
  SimTK Exception thrown at polygonalmesh.cpp:340:
  Error detected by Simbody method PolygonalMesh::loadVtpFile(): Only format="ascii" is supported for .vtp file DataArray elements, got format="appended" for Points DataArray.
  (Required condition 'pointData.getRequiredAttributeValue("format") == "ascii"' was not met.)

I am not sure if this is what's causing the issue since they are different models than Laurent was using.

Thanks!
Maria

[Ayman Habib]

Hi Maria,

The visualization issue is definitely unrelated to Scaling. The reason some vtp files show in 3.3 and not in 4.0 is that these were binary vtp (rarther than ASCII) which were better handled by the old visualizer. There are tools to convert binary to ASCII vtp file formats and the 4.0 distribution has only ASCII vtp files but in general these vtp files are used exclusively for visualization and do not affect anything else.

Possible reasons for the scaling issue you report:
- Permissions (not being able to write intermediate or result files).
- Naming: we enforce strict uniqueness of names, which are enforced during model/files reading. Please check the GUI panels in 3.3 and 4.0 to make sure the files/inputs to scaling are the same. Check the messages window to see if Body or Marker was renamed.

Let us know what you find out.
-Ayman

[Maria Fox]

Ayman,

Thanks for the quick reply. I don't think there are any issues with permissions. In part 2 of scaling, it is able to write the output scale file (Sub_ScaleSet_Applied.xml) to the same directory with no issues. I have checked the paths in each part of the scale file and they appear to be okay as well.

When I load the Lai 2017 model, the only message that appears is the error when trying to load the hat_ribs_scap.vtp file. It does not look like anything is being renamed. There are no duplicate marker names, and I don't see any messages about Body or Marker being renamed.

I'm not sure what else could be the problem given that I'm using the exact same models and setup files to run scaling in both 3.3 and 4.0 and don't have any obvious errors.

Thanks!

[Juan-Manuel Belda-Lois]

I'm experiencing the same problem. I'm still using Opensim 3.3 because I'm not able of performing any scaling neither in Opensim 4.0 or OpenSim 4.1.

I am using the gait2354 model with the a set of markers I use in OpenSim 3.3. I have re-defined a new set of markers (instead of importing) but I am not able of performing the scaling.

[Ayman Habib]

Hello,

If you can perform the model scaling in Tutorial 3 then we can rule this out as model specific issue, in this case I'd suggest you file a bug and attach the model and all the files necessary to reproduce your scaling issue and we'll investigate (could be permissions, special characters in marker names or file/path names, or just syntax in xml files). Also keep in mind that 4.0+ enforces strict naming convention so pay special attention to messages that show up when you open the model first time in 4.0+.

Hope this helps,
-Ayman

[Juan-Manuel Belda-Lois]

I'm still having problems with this issue.
I am able to scale the model with OpenSim 3.3 but not with OpenSim4.0

This is the message of output:

Loaded marker file D:\Formacion\DisennoOP\PracticaMarcha\escalado.trc (26 markers, 290 frames)
Measurement 'pelvis'
pair 0 (R.ASIS, L.ASIS): model = 0.256, experimental = 0.216462
overall scale factor = 0.845554
Measurement 'muslo'
pair 0 (R.Thigh.Upper, R.Knee.Lat): model = 0.375303, experimental = 0.418377
pair 1 (L.Thigh.Upper, L.Knee.Lat): model = 0.375303, experimental = 0.392455
pair 2 (R.Knee.Lat, R.Knee.Med): model = 0.1, experimental = 0.128565
pair 3 (L.Knee.Lat, L.Knee.Med): model = 0.1, experimental = 0.129113
overall scale factor = 1.18431
Measurement 'pierna'
pair 0 (R.Ankle.Lat, R.Ankle.Med): model = 0.0940821, experimental = 0.0968878
pair 1 (L.Ankle.Lat, L.Ankle.Med): model = 0.0940821, experimental = 0.0994085
pair 2 (R.Shank.Upper, R.Ankle.Lat): model = 0.345158, experimental = 0.346773
pair 3 (L.Shank.Upper, L.Ankle.Lat): model = 0.345158, experimental = 0.361664
overall scale factor = 1.03473
Measurement 'Retropie'
pair 0 (R.Heel, R.Midfoot.Lat): model = 0.174642, experimental = 0.201262
pair 1 (L.Heel, L.Midfoot.Lat): model = 0.174642, experimental = 0.199151
pair 2 (R.Midfoot.Lat, R.Midfoot.Sup): model = 0.0761577, experimental = 0.10708
pair 3 (L.Midfoot.Lat, L.Midfoot.Sup): model = 0.0761577, experimental = 0.103299
overall scale factor = 1.26379
Measurement 'Torso'
pair 0 (L.Acromium, R.Acromium): model = 0.297378, experimental = 0.313319
pair 1 (R.ASIS, L.ASIS): model = 0.256, experimental = 0.216462
pair 2 (Top.Head, Sternum): model = 0.281603, experimental = 0.276309
overall scale factor = 0.960121
Loaded marker file D:\Formacion\DisennoOP\PracticaMarcha\escalado.trc (26 markers, 290 frames)

Step 2: Scaling generic model
Loaded marker file D:\Formacion\DisennoOP\PracticaMarcha\escalado.trc (26 markers, 290 frames)
Measurement 'pelvis'
pair 0 (R.ASIS, L.ASIS): model = 0.256, experimental = 0.216462
overall scale factor = 0.845554
Measurement 'muslo'
pair 0 (R.Thigh.Upper, R.Knee.Lat): model = 0.375303, experimental = 0.418377
pair 1 (L.Thigh.Upper, L.Knee.Lat): model = 0.375303, experimental = 0.392455
pair 2 (R.Knee.Lat, R.Knee.Med): model = 0.1, experimental = 0.128565
pair 3 (L.Knee.Lat, L.Knee.Med): model = 0.1, experimental = 0.129113
overall scale factor = 1.18431
Measurement 'pierna'
pair 0 (R.Ankle.Lat, R.Ankle.Med): model = 0.0940821, experimental = 0.0968878
pair 1 (L.Ankle.Lat, L.Ankle.Med): model = 0.0940821, experimental = 0.0994085
pair 2 (R.Shank.Upper, R.Ankle.Lat): model = 0.345158, experimental = 0.346773
pair 3 (L.Shank.Upper, L.Ankle.Lat): model = 0.345158, experimental = 0.361664
overall scale factor = 1.03473
Measurement 'Retropie'
pair 0 (R.Heel, R.Midfoot.Lat): model = 0.174642, experimental = 0.201262
pair 1 (L.Heel, L.Midfoot.Lat): model = 0.174642, experimental = 0.199151
pair 2 (R.Midfoot.Lat, R.Midfoot.Sup): model = 0.0761577, experimental = 0.10708
pair 3 (L.Midfoot.Lat, L.Midfoot.Sup): model = 0.0761577, experimental = 0.103299
overall scale factor = 1.26379
Measurement 'Torso'
pair 0 (L.Acromium, R.Acromium): model = 0.297378, experimental = 0.313319
pair 1 (R.ASIS, L.ASIS): model = 0.256, experimental = 0.216462
pair 2 (Top.Head, Sternum): model = 0.281603, experimental = 0.276309
overall scale factor = 0.960121

Step 3: Placing markers on model

[Juan-Manuel Belda-Lois]

Besides to the previous post. This is the message of the error

java.lang.RuntimeException: invalid vector<T> subscript
at org.opensim.modeling.opensimActuatorsAnalysesToolsJNI.OpenSimContext_processModelMarkerPlacer__SWIG_0(Native Method)
at org.opensim.modeling.OpenSimContext.processModelMarkerPlacer(OpenSimContext.java:305)
at org.opensim.tracking.ScaleToolModel$ScaleToolWorker.construct(ScaleToolModel.java:350)
at org.opensim.swingui.SwingWorker$2.run(SwingWorker.java:130)
[catch] at java.lang.Thread.run(Thread.java:748)

[Juan-Manuel Belda-Lois]

Finally done it.

The problem was in the .trc in OpenSim 4.x the processor of the .trc seems to be far more intransigent with sintaxis than the OpenSim 3.x versions.