Open utterances-bot opened 3 years ago
hello sir, ran your tutorial in unity and extremely thankfull for the detailed explanation of the codes wanted to know if i can connect the x y outputs to a progress bar to achieve a particular body pose.
Hi @rawde94, If you have a list of reference (x,y) key point coordinates for the target pose, you could use the Vector2.Distance method to calculate the distance between the estimated and reference key point locations. If the distance for a key point is small enough, it could be marked as being in the correct position.
The progress bar could then be updated based on the number of key points in the correct position.
This approach would require that the reference key point coordinates and the estimated key point coordinates be for the same image dimensions.
Also, I recently released an updated version of this tutorial.
Barracuda PoseNet Tutorial Pt. 8 | Christian Mills
This post covers how to handle video input with different aspect ratios.
https://christianjmills.com/Barracuda-PoseNet-Tutorial-8/