values for the ankle

New project for OpenCap, which is a new software package to estimate 3D human movement dynamics from smartphone videos. OpenCap strongly relies on OpenSim.
POST REPLY
User avatar
Maximilian Puhr
Posts: 13
Joined: Fri Nov 17, 2023 3:40 am

values for the ankle

Post by Maximilian Puhr » Fri Feb 02, 2024 6:57 am

Hello,

In my masters thesis I write about the influence of clothing conditions on OpenCap and I have a problem regarding values for the ankle.

Attached are two images showing ankle kinematics in the sagittal plane. Noticeably, there is a lack of plantarflexion in swing for normal walking in the right image. In experiments from a year ago (left), this appeared normal. Left: Data from a year ago for normal walking (OpenCap in pink) - right (Data for normal walking with only OpenCap and two clothing conditions).

Could it be that there is an issue with the model right now, which generates "wrong" ankle-values?

Best regards,
Maximilian Puhr
Attachments
Ankle.png
Ankle.png (233.11 KiB) Viewed 885 times

User avatar
Antoine Falisse
Posts: 438
Joined: Wed Jan 07, 2015 2:21 am

Re: values for the ankle

Post by Antoine Falisse » Fri Feb 02, 2024 6:12 pm

Do you mean you had good results last year but bad results this year?
Is that with the same data?

User avatar
Maximilian Puhr
Posts: 13
Joined: Fri Nov 17, 2023 3:40 am

Re: values for the ankle

Post by Maximilian Puhr » Sun Feb 04, 2024 7:02 am

Hi and thank you for your response!

Yes, I mean last year were the curves very similar to curves measured by other motion-capturing systems while assessing physiological gait. This year, the curves differ quite a lot to other systems, as you can see in the pictures.

The data are different, but both trials measured physiological gait.

Could it be that something in the algorithms for the ankle has changed?

Best regards

User avatar
Antoine Falisse
Posts: 438
Joined: Wed Jan 07, 2015 2:21 am

Re: values for the ankle

Post by Antoine Falisse » Sun Feb 04, 2024 4:59 pm

Hi, you could try reprocessing the data using the previous augmenter model, which is more specific to gait, to see whether you get more expected results:

1. Step 1: change the session metadata by setting augmentermodel v0.2 instead of v0.3: https://github.com/stanfordnmbl/opencap ... ata.py#L30

2. Step 2: reprocess the data: https://github.com/stanfordnmbl/opencap ... essions.py

This video describes the process: https://www.youtube.com/watch?v=Mk3KPo7 ... e=youtu.be

User avatar
Maximilian Puhr
Posts: 13
Joined: Fri Nov 17, 2023 3:40 am

Re: values for the ankle

Post by Maximilian Puhr » Tue Feb 06, 2024 10:29 am

Thank you very much for your effort! I am going to try this approach

Best regards

User avatar
Antoine Falisse
Posts: 438
Joined: Wed Jan 07, 2015 2:21 am

Re: values for the ankle

Post by Antoine Falisse » Wed Feb 07, 2024 10:39 am

Keep us posted

POST REPLY