xkunwu/depth-hand

running on live camera

ankurhanda opened this issue · 3 comments

ankurhanda commented

Is there any simple example which shows how to run the code for depth images coming from a live camera. something like

img = camera.get_image()

out = network.run(img)

display(out)

I tried to look into it but I can't find such an example easily.

ankurhanda commented

Also, where can I find the model.ckpt.meta - is there any pretrained model? I looked into running camera.capture python script to see if I can get it running live on my realsense camera but I don't find the pre-trained model anywhere...

ankurhanda commented

rsync-proj.sh didn't work for me properly. How do you run rsync-proj.sh file? it seems to need arguments etc. Can you give an example please?

xkunwu commented

Hi,

  • please read both "README" for estimation and detection first.
  • camera.capture contains capture code.
  • pretrained model can be downloaded, and instructions are in the "README".
  • you can ignore rsync-proj.sh.