Closed LimitlessGrey closed 1 year ago
Do you think that method will be fast enough, I mean not leave the showing of images very slow?
No idea. The code itself runs quite well alone.. Im not sure if it would do the same when implemented together with what we have already. We can try it :)
Okay seems like a good next step for now.
Just did tracking of the body where it recognizes it as a skeleton (hands included). can you help me pass it to a class in functions @andrefdre, please?
Forgot to add: this holistic model to track the body moving requires to install mediapipe via pip : !pip install mediapip (! may or may not be necessary)
I found a good method of recognizing bodies that works very well and I was able to implement it replacing the Haar Cascade. I based my code on this repository https://github.com/Kazuhito00/Person-Detection-using-RaspberryPi-CPU/blob/main/README_EN.md.
https://github.com/Its-Triggy/Joint_Track/blob/master/jointTrack.py
Found this (apparently ineficient) code to perform body tracking. It seems it uses a system of joints and branches and recognizes the body as such