Hello,
I can see cumulative effect of noise after 5-6 seconds. Could you help me with more information about the noise and how to deal with it?
Best Regards,
Riva Karia
Noticing cumulative effect of noise
-
Matt Petrucci
- Posts: 166
- Joined: Fri Feb 24, 2012 11:49 am
Re: Noticing cumulative effect of noise
Hi Riva,
Can you be a bit more specific about what activity you are trying to collect, and what you are seeing in the kinematics?
Can you be a bit more specific about what activity you are trying to collect, and what you are seeing in the kinematics?
-
Riva Karia
- Posts: 5
- Joined: Fri Jun 21, 2024 3:50 pm
Re: Noticing cumulative effect of noise
Hi Matt,
The tasks i am performing are Sit to stand and walking tasks. I have noticed that after a few seconds there is accumulation of noise and is visible in the graphs.
Best Regards,
Riva Karia
The tasks i am performing are Sit to stand and walking tasks. I have noticed that after a few seconds there is accumulation of noise and is visible in the graphs.
Best Regards,
Riva Karia
-
Matt Petrucci
- Posts: 166
- Joined: Fri Feb 24, 2012 11:49 am
Re: Noticing cumulative effect of noise
Hi Riva,
There could be several causes for what you are describing. If there is a specific session that demonstrates this you can pass it along to me.
But here are a few things:
-Camera sync - if you are doing cyclic movements like gait or sit-to-stand, the algorithm can sometimes have a hard time synchronizing the videos. You can start the trial with a hand punch over head to help the algorithm synchronize.
- Camera angles - If you are in a two camera setup, you want to have them as close to 45/45 to the participant as possible to get more 3D information. A purely frontal or sagittal camera can sometimes throw off the algorithm, especially when there are occlusions.
- FOV - Make sure your participants are clearly visible at all times. Sometimes if they are too small in the FOV, the pose estimation can have a hard time seeing specific body landmarks, adding to noise.
These tips and others are on the best practices page. For setting up your FOV, I would suggest watching the beginning of the tutorial video on that page.
Hope these help,
There could be several causes for what you are describing. If there is a specific session that demonstrates this you can pass it along to me.
But here are a few things:
-Camera sync - if you are doing cyclic movements like gait or sit-to-stand, the algorithm can sometimes have a hard time synchronizing the videos. You can start the trial with a hand punch over head to help the algorithm synchronize.
- Camera angles - If you are in a two camera setup, you want to have them as close to 45/45 to the participant as possible to get more 3D information. A purely frontal or sagittal camera can sometimes throw off the algorithm, especially when there are occlusions.
- FOV - Make sure your participants are clearly visible at all times. Sometimes if they are too small in the FOV, the pose estimation can have a hard time seeing specific body landmarks, adding to noise.
These tips and others are on the best practices page. For setting up your FOV, I would suggest watching the beginning of the tutorial video on that page.
Hope these help,