Keypoints measurement

Hello everyone,
I’ve been reviewing the keypoints of Nuitrack and I’d like to delve a bit deeper into them. I’m particularly interested in the following: neck, collar, torso, and waist. I’d like to know how Nuitrack locates these keypoints. For example, we know that the torso lies between the waist and the collar, but how does it specifically determine the distance it should be from each of these? Does it adjust according to each person’s anatomy, aiming to always be on a specific vertebra? Or does it follow a scale? For example, if we consider the waist as 0% and the collar as 100%, should the torso always be at 30%?

Where can I read more about this topic?

Thank you.

Hi @daniela

Or does it follow a scale? For example, if we consider the waist as 0% and the collar as 100%, should the torso always be at 30%?

You described more or less correct model of how it could be thought of.
Specifically torso/collar/waist are a “virtual joints” which is calculated from other joints.

We are working on biomechanical internal model which will bind most joints to the specific precise “locations” in human body (correspond to your first guess - “aiming to always be on a specific vertebra”) so these joints definitions planned to be re-defined/refined with more precision.

Hi @daniela ,

We’ll add you in the waiting list for the “biomechanical model” if you wish.
Please also let us know if there anything else in Nuitrack experience we could help with.

Hi @TAG ,

Thank you for the help. I’m already working in it, but thank you for your help. However, I would really appreciate the possibility of recording the coordinates of two people at in the same trial (as I said in another post).

Thank you

We’re thinking on further development of this tool, but can’t see a particular use-cases for it.

If you could help us with some details on your use-case / project - please feel free to message us through a feedback form on https://nuitrack.com

Hi @TAG ,
I couldn’t find the feedback form, sorry.

Do you mean particular use-cases for the recorder itself? Or for it being able to record more than one person?

I think the recorder would be really useful for people that is not expert in programming which would at the end increase the community that uses Nuitrack.

And why recording more than one person? I personally choose nuitrack for my project because of two reasons: The keypoints Nuitrack provides are really useful for computing moments/compression forces in the L5-S1 region of the back. The second reason was that Nuitrack can track more than one person at the same time. If you compare it with the standard that is the motion capture system, the motion capture system can track only one person (unless you have an incredibly huge amount of markers), and it is necessary to be in a lab environment. On the other side, Nuitrack could be used in any place and if it was possible to get data for more than one person at the same time, the number of trials you could take would be incredibly high.

Thanks for all the help.

Thank you for all the details. Really appreciate it.

I couldn’t find the feedback form, sorry.

Sorry for confusion, please refer to the answer at:

I personally choose nuitrack for my project because of two reasons

Could you kindly clarify - is it a commercial or research project ? And what are primary expected outcome - research paper, pre-production prototype, MVP of a new product, etc ? This information would help us to figure out on how to proceed with the current recorder prototype.
Please feel free to choose where it’s more convenient to continue discussion - at this topic or through a custom project inquiry (through an email).