Question about recording time

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
Wisavaporn Niamsang
Posts: 5
Joined: Tue Nov 01, 2022 8:50 pm

Question about recording time

Post by Wisavaporn Niamsang » Wed Jan 03, 2024 8:22 am

Hi,
For collecting and analyzing data on human gait and movement, OpenCap is incredibly useful, which is why I truly appreciate it. As a part of my research, I am interested and have taken OpenCap. However, I have a question. I would like to know that when I record data and I use the farm rate at 120fps then I try to record the data for around 10-20 seconds. But the data that I got after that is just 3-4 seconds. I'm not sure how it happened and how can I solve that problem.

Thank you

User avatar
Antoine Falisse
Posts: 431
Joined: Wed Jan 07, 2015 2:21 am

Re: Question about recording time

Post by Antoine Falisse » Wed Jan 03, 2024 9:44 am

Hi,
The data is trimmed if the person is not visible by at least two cameras in the beginning and end of the trial.
Best,
Antoine

User avatar
Wisavaporn Niamsang
Posts: 5
Joined: Tue Nov 01, 2022 8:50 pm

Re: Question about recording time

Post by Wisavaporn Niamsang » Wed Jan 03, 2024 10:04 am

Thank you for your good answer.

I'm pretty sure both cameras can see and capture the whole body of a patient at the beginning and end of the trial. But I'm not sure if it might have something wrong to do with my patients who use prosthetics during the data collection process.

Thank you very much

User avatar
Scott Uhlrich
Posts: 124
Joined: Tue Jan 20, 2015 4:29 pm

Re: Question about recording time

Post by Scott Uhlrich » Wed Jan 03, 2024 10:18 am

If at least 2 cameras cannot see all keypoints on the person, it will assume they have left the frame and trim the trial. If this problem is showing up with prosthesis users, my guess is that the pose detection algorithm is not recognizing some of the joints on the prosthesis. The more like a biological limb you can make it look (eg wearing pants), the better the pose estimatoin model will likely do.

POST REPLY