Detecting people in a room with ML

vvvv gamma question
1

So I had a look VL.YOLO.GPU and tried to install the Cuda files but somehow I didn’t get it to work since I couldn’t figure out where I could install the files regarding YoloApi and cudnn64_7.dll.
Now my question is, is this the latest and best method to track people (around 15) or is there in the mean time stuff that works even better and is easier to implement?

Kind regards,
Y

2

Some time ago, when there was no YOLO for VL, I made this patch below on OpenCV (the patch itself is lost, probably). I think it’s the best solution, but it will take a bit of figuring out the libraries and where to put them. Where did you download VL.YOLO.GPU from?

3

@yar There was this wip thread by @bjoern

VL.Yolo.GPU - wip - Forum (vvvv.org)

1 Like
4

I also found VL.Mediapipe which works quiet well:)

1 Like
5

At the moment I am sending texture to ndi, ndi to webcam, and from there using yolov8/11 to do the inference, and send back the result via udp in json format. Not pretty, but it do the job